[PROB] TouchPro GPS lag problems

[Sleuth255]

Why don't you try a route simulation? I'm waiting for the results :D.

I've tried on igo and routes simulation are ok (not perfect smooth, maybe due poor graphics drivers or software.. but it's definitely better than 'real-time routes')

Ok, good point. Its probably due to the greater speed lag issue on the fuze then.

[noellenchris]

HKLM\Drivers\USB\FunctionDrivers\Serial_Class\Conf ig

does "EnableFlowHard" & "EnableFlowSoft" work just like modems? If so, both were set to 0. I changed the Hard to 1. Who knows if it actually is doing anything for me or not.

I was using:

HKLM\System\CurrentControlSet\GPS Intermediate Driver\Multiplexer

MaxBufferSize = 256 (I changed this for my next test to 128)

HKLM\System\CurrentControlSet\GPS Intermediate Driver\Drivers

InputBufferSize = 256
OutputBufferSize = 128
SleepOnNoData = 1000

HKLM\System\CurrentControlSet\GPS Intermediate Driver\Drivers\GPSOneDevice

PollInterval = 250

My Log files are set to 0.

My Performance settings are:

File sys cache = enabled
File sys cache size = Auto (going to change to 8mb)
File sys filter cache = disabled (going to change to 131072)
Glyph cache = 32KB (going to change to 128kb)

Will test again with TT7 using internal GPS tommorow.

Finally making some progress:)

[Sleuth255]

We've not found any evidence that InputBufferSize and OutputBufferSize in the Drivers key is used by the GPSID. Da_G noticed this first and I can confirm that I didn't see them when I had the driver disassembled in IDA. MSDN doesn't mention these values here either. Nevertheless, they persist not only as OEM settings in my fuze, but they're in my Kaiser as well. Maybe something other than the GPSID uses them. FWIW, my Fuze is set to 128 for both of these.

To me, the key values seem to be MaxBufferSize in Multiplexer along with PollInterval in GPSOneDevice. Mine are 256 and 200 respectively.

[Da_G]

Run RegLogger when you open up a GPS connection (love this program :))

You'll see the GPSID access those values, but not under this key where all the ROM's have it by default:

[HKEY_LOCAL_MACHINE\System\CurrentControlSet\GPS Intermediate Driver\Drivers]

Instead it looks under this key:

[HKEY_LOCAL_MACHINE\System\CurrentControlSet\GPS Intermediate Driver\Drivers\GPSOneDevice]

and reads those values. this is the microsoft standard as per this page. (http://msdn.microsoft.com/en-us/library/ms889971.aspx)

Our GPSID is using the POLL interface rather than COMM, so it's not even supposed to be utilizing those keys. (they are only used for the COMM interface) - nonetheless, the GPSID can clearly be seen reading these keys by reglogger. However, again, they are not read from the location the tweaks recommend you set it.

[Sleuth255]

That would explain why I didn't see them referenced for the Drivers key. Maybe something else besides GPSID references them.

[Da_G]

Take a look at offset C18 in a hex editor (gpsid_htc 2.03 from diamond rom)

[mindfrost82]

Any idea about the settings under Multiplexer\ActiveDevice?

I have:

Context = 305414496
Dll = GPSID_HTC.dll
Flags = 2
Index = 4
Keep = 1
Prefix = COM

[mindfrost82]

@Sleuth: After installing your cab and shutting off my phone I was able to lock onto 7 sats in my living room within about 15 seconds.

I did change my Multiplexer Buffer to 128.

I've never been able to get a lock indoors with my Fuze (but I usually could with my Tilt).

[Sleuth255]

@Da_G: Yes they are there. Sorry that I wasn't being clear. I was looking for value references under Drivers in IDA and I didn't see them.

I guess my point was not that GPSID uses these values in GPSOneDevice, but rather that something else may use values having the same names in Drivers. It seems odd to me that the error would persist through two generations of hardware...

[Sleuth255]

@Sleuth: After installing your cab and shutting off my phone I was able to lock onto 7 sats in my living room within about 15 seconds.

I did change my Multiplexer Buffer to 128.

I've never been able to get a lock indoors with my Fuze (but I usually could with my Tilt).

It's that blackstone radio. I'm seeing a whole different acquisition sequence in visualgpsce since I installed it. Much more like my Tilt in that it lights up slots for birds that are in the sky rather than always showing "16 visible" with 8 filled with random PRN's and the others empty like my fuze used to show.

Also: it appears that GPSID sets up those values itself in the ActiveDevice key.

[mindfrost82]

It's that blackstone radio. I'm seeing a whole different acquisition sequence in visualgpsce since I installed it. Much more like my Tilt in that it lights up slots for birds that are in the sky rather than showing "16 visable" with 16 other empty slots like my fuze used to show.
Before installing your cab I never got a lock inside, even with the Blackstone radio.

Tried the new Xperia radio yet? I haven't, but I might.

[Da_G]

Here's a link to the MSDN reference (http://msdn.microsoft.com/en-us/library/bb202042.aspx) for that key. this page (http://msdn.microsoft.com/en-us/library/aa926416.aspx) also has more info on those keys.

From what I know, flags = 2 tells WinMo to use LoadLibrary instead of LoadDriver, which means the code can be paged out when not in use. That shouldn't affect anything, but you can set flags = 0 to disable that behavior.

Keep = 1 tells the system not to unload the dll. Prefix COM and Index 4 means to expose the device as COM4:, and DLL=gpsid_htc.dll tells the system which dll provides the multiplexer service.

The Context value gets passed along to the intialization routine of GPSID_HTC.dll, but I don't know what it does with it.

Sleuth: Surprising, yes, out of the ordinary, not so much, I think.. i've fixed quite a few typos in HTC's packages.. there's one that persists through the newest version even now for example, HTC Streaming Media references the following key:


[HKEY_LOCAL_MACHINE\System\Inbox\SICInt\SIPProtocol s]
"rtsp"=dword:00000000


There's a typo in \Inbox\SICInt\, it should be \Inbox\SIClnt (that's an ell not an eye), so the registry entry has no effect as-is.

I can think of at least 10 similar registry errors i've fixed..

So, I wouldn't be terribly surprised if it wasn't read somewhere else. I do however know an easy way to find out.. EMEditor's awesome search-in-files function can search UTF-16, so it should be a matter of seconds to answer that question.. /me searches.

aaand that's a no. The only file in \OEM that references this key is GPSID_HTC.dll. So I feel pretty confident in saying that it's a persistant registry error in HTC's platform builder template..

[X2D]

I have to security unlock my phone to get the blackstone radio right? :(

[mindfrost82]

I have to security unlock my phone to get the blackstone radio right? :(
Yup, you do, but its been worth it for me because I've loved the call quality of the Blackstone radios compared to the Raphael radios I've tried.

[noellenchris]

I have to security unlock my phone to get the blackstone radio right? :(

Yes , you have to security unlock first.. I did it yesterday and I'm unlocked..works nice

[Sleuth255]

@X2D: yes :(.
@Da_G: that settles that one then. I'll let someone else play with them in GPSOneDevice (although they shouldn't have an effect on POLL interfaces anyway)

[Sleuth255]

Yup, you do, but its been worth it for me because I've loved the call quality of the Blackstone radios compared to the Raphael radios I've tried.

I noticed that myself!

[noellenchris]

I noticed that myself!

Man, I've never noticed any difference in battery life with radios on my previous phones...but I'll add that the BS radio is noticeably better. This one is like the energizer bunny...lol

[Sleuth255]

Before installing your cab I never got a lock inside, even with the Blackstone radio.


hmmm... maybe some combination of both then. I have so many tweaks in the GPSID key that I can't remember them all :rolleyes: I'm gonna stick that cab into my UC I think.

edit: all joking aside, I wouldn't recommend anyone putting that cab into UC. It's for debugging only and has device specific information that could very well break the GPS settings on another device or even a subsequent generation of ROMs for this one.

[mindfrost82]

hmmm... maybe some combination of both then. I have so many tweaks in the GPSID key that I can't remember them all :rolleyes: I'm gonna stick that cab into my UC I think.
Well next time I go driving I'll test all my lags and see what happens :)

[Da_G]

Has anyone tried just nuking the HTC GPSID and using the Microsoft GPSID?

I know I personally haven't, but i'm thinking it's a logical next step to eliminating gpsid_htc.dll as a source of lag.. if no one's tried it i'll whip up a rom right quick.

[Sleuth255]

I was going to throw in the one from my Tilt for testing but I can't reproduce position lag on my device.

[Da_G]

OK, I think we need to step beyond casual testing at this point, then.

I'll cook up a ROM with the bare neccessities to get GPS working (gpsdriver.dll, microsoft gpsid, no agps or quickgps)

I'll follow the microsoft guidelines for setting up the poll interface, with no extra keys.

Assuming it works, that should at least help to eliminate some variables in the equation, here.

[Sleuth255]

Do you have position lag currently Da_G?

[X2D]

I was going to throw in the one from my Tilt for testing but I can't reproduce position lag on my device.

Salt in the wound.. but thanks for the help I wanna do everything I can to test the crap out of this GPS until we find something out. New ROM = fail, still broken using it for tomorrow to do one more test before switching back to Da_G's but I might try the stock AT&T Radio again just for s&g's.

[Sleuth255]

Salt was unintentional. Only trying to say that its hard to fix something when I can't reproduce the problem. Maybe more GPS satellites fly by Milwaukee WI than by everywhere else.

Too bad that nobody with position lag issues has posted a reg file containing their GPSID keys...

[Da_G]

Yes, I have both "Position Lag" and "Speed Lag"

It seems to vary quite a bit. Sometimes a soft reset will clear it right up immediately..

Sometimes it goes away, sometimes it comes back real quickly. Even without changing anything at all, which makes it quite difficult to diagnose reliably..

What I really want is one of the GPS signal generators they use to test these devices. It creates a fake 'sky' with fake sats and everything, you can have it report any position/time you like.. that would be perfect to debug something like this. Too bad i'm sure it costs more than my house.. lol

Prepping to flash the gpsid-neutered rom now. Let's see if microsoft's gpsid performs any better.. if it works!

[Sleuth255]

Well, if position lag is caused by nmea sentence buffering then the GPSID should be the responsible bugger. We could also try the simple RIL shim that was written for hermes. It sends AT@StartGPS when its com port is opened and AT@StopGPS when the port is closed. In between it streams NMEA sentences that it receives. Pretty simple.

If anyone wants to play with it, here's the link to that development thread:

Hermes GPS Driver (http://forum.xda-developers.com/showthread.php?t=294113)

[Da_G]

Ok, preliminary testing:

HTC GPS Tool doesn't seem to work at all with the microsoft supplied GPSID.

I've got GPS Test loaded trying to get a fix, but without QuickGPS I think it's going to take forever in this room.. it's at least showing sats and whatnot so that shows the microsoft supplied gpsid does work.

[Ninja1]

Could somebody experiencing position lag post a reg containing their HKLM\System\CurrentControlSet\GPS Intermediate Driver key along with the sub keys/values.

use dotFred and just export the key then attach. Thx!

Sorry for the delay, I am trying to get Win7 to install on a couple of boxes. Here is my reg file, using latest DaG ROM with iGuidance and I have both position and speed lag.

Currently running with the ROM defaults for all GPS related settings...hopefully this will help me determine what to tweak.

[petard]

AT&T FUZE Stock ROM
2-4 seconds positional lag

[Da_G]

OK, the ROM tests good, uploading it now.

What it is: 2.07-Pre version of my ROM, with everything possible stripped as far as GPS goes.

AGPS is disabled. Re-Aiding is disabled. GPSID_HTC.dll is removed and GPSID.dll from microsoft is used. So the only HTC code in play here is the RIL and gpsdriver.dll.

HTC GPS Tool doesn't work, but GPS Test 1.04 and google maps, etc all work fine.

Let's see where this takes us!

[petard]

Sleuth, have you tried going back to a stock AT&T config to see if your FUZE has positional lag? I really appreciate all the work you guys are doing. I hope you are able to solve this. It is amazing how HTC continues to release such buggy hardware. I hope they get off of Qualcomm and start releasing good, working hardware soon.

[Da_G]

Here's the link to the ROM with microsoft's GPSID:

http://rapidshare.com/files/181978641/2.07-pre-standalone-gps-mode.rar

Anyone interested in really getting this fixed, who has some time to spare, let's get some testing in with this ROM, so we can eliminate GPSID_HTC.dll as the source of the problem.. or heck, maybe the problem isn't even there in this ROM since we aren't using gpsid_htc.dll. Test, test, test! :)

[Ninja1]

Rapidshare says that I have exceeded my limit, so I am gonna have to try this in the am (PST).

[X2D]

Salt was unintentional. Only trying to say that its hard to fix something when I can't reproduce the problem. Maybe more GPS satellites fly by Milwaukee WI than by everywhere else.

Too bad that nobody with position lag issues has posted a reg file containing their GPSID keys...

Working on it! (joking about the salt, the help really is appreciated)

Could you attach the app or should I just do a handy dandy search =P~

Been working on work project haven't had time to do phone stuff. I used your CAB would that change my GPSID key?


Here's the link to the ROM with microsoft's GPSID:

http://rapidshare.com/files/181978641/2.07-pre-standalone-gps-mode.rar

Anyone interested in really getting this fixed, who has some time to spare, let's get some testing in with this ROM, so we can eliminate GPSID_HTC.dll as the source of the problem.. or heck, maybe the problem isn't even there in this ROM since we aren't using gpsid_htc.dll. Test, test, test! :)


Will download this tonight and give it a shot with TT7, iGO8, OCN8, Garmin XT

[maw]

Okay.... I unlocked my device and flashed the BS radio .25 on it. I'm going to check if I get the same responses as you guys here!

Keeping my fingers crossed....

[Sleuth255]

DotFred's Taskmanager is Here (http://www.dotfred.net/TaskMgr.htm)

Thx for the ROM Da_G; this'll really help eliminate or prove GPSID. What's in the GPSID keys for this version? (edit: never mind, I'll find out myself)

Ok, so who's planning on testing Da_G's MS_GPSID ROM besides X2D?

There's no warm start capabilities in this ROM so load up GPS Test or Visualgpsce, turn sleep mode off and leave it in a location that has a clear view of the sky. The first cold fix should take around 10 minutes. After that, you will be able to hot start. Satellite windows are a couple of hours which means that you have that much time between tests before you'll have to Cold Start again.

[Sleuth255]

The two of you who uploaded GPSID keys all have a few things in common. I'm going to make a cab for you to try.

Edit: Here's the cab. Let me know if it affects your position lag.

[maw]

I tested the BS radio.

First the GPS locked but then it was going off and on for a couple times. When it got a steady lock then it stayed locked. Did any of you have this the first time with the BS radio? I hope it stays locked now for good....

Second test I did I called I went outside and called the house phone. The sound quality was superb! No electric kind of static and crystal clear! Massive improvement here!

Now I'm going to see what it does with battery life.... I had very good battery life with the RAPH HTC radio .11.... I hope this one is indeed the energizer bunny! Will let you know more when in a couple hours.... or maybe in the morning.

[Sleuth255]

The lock effect that you noted happens with mine too. It flips a few times before getting a solid lock. However the total time to a solid lock is so much less than with any other radio I tried that I have no issues with it.

[maw]

Okay cool... still kinda weird though. Anyway, so far battery level dropped 2% with navigation software running 5 minutes and making 1 phone call of 2 minutes.

[noellenchris]

DotFred's Taskmanager is Here (http://www.dotfred.net/TaskMgr.htm)

I've been using that task manager for years since my Wizard! I love it. I just d/l the 3.1 version since mine was 2.9. It really does so much! :)

[hopper13]

Among the registry values it uses, try checking here:

HKLM\System\CurrentControlset\GPS Intermediate Driver\Multiplexer

You all do have MaxBufferSize set to 256 right?

Under that key, I have only the following listings:

\ActiveDevice
+Add value+
DriverInterface
---------
I have no listing for a MaxBufferSize on my Fuze!

[Sleuth255]

Yep, most people don't. Try the cab I attached to this (http://forum.xda-developers.com/showpost.php?p=3152426&postcount=539) post.

After installing, reset your handheld by powering off then back on. Do not just press the reset button. This will force a write-back of the registry changes and will also restart GPSID with the new settings.

[hopper13]

The two of you who uploaded GPSID keys all have a few things in common. I'm going to make a cab for you to try.

Edit: Here's the cab. Let me know if it affects your position lag.

Thanks. Doing a soft-reset BEFORE installing. I will install, power off and on, go for test ride and report back.

Prior to this, my position lag seems to run between 1.5 and 2.5 seconds depending upon conditions. My speed lag is generally up to 5 seconds and time to lock runs about 5 minutes after a reset with a fresh quickgps and then under 20 seconds for future TomTom 7.9 starts.

[X2D]

I won't be able to test flashing results until after 4-5pm cst as soon as I go home I'm going to flash Da_G's latest with everything stripped and try it on a set of dif nav systems.

A) First test will be Da_G's newest stock nothing else
B) Second test will be with the additional Cab that you first posted (Sleuth)

Will try to expand the tests by including assisted GPS in the mix if his ROM allows me. The second cab you posted, what exactly does that change?

[Sleuth255]

@hopper13: This cab is an attempt to lower position lag only. I don't think there's much that can be done for speed lag. Warm start times are dependent on the radio you are running and how you have Quick GPS/aGPS set.

If you are in the US, then I recommend these settings to minimize WarmStart times:

Turn aGPS off (in the system tab of settings).
Set QuickGPS to obtain updates automatically then have it download current data.
Flash the BlackStone radio (see my sig) if you can get security unlocked.


@X2D: the cab primarily plays with MaxBufferSize in Multiplexer, PollInterval in GPSOneDevice, and SleepOnNoData in Drivers. It also turns off logging. They are the settings I have on my Fuze.

[aevh]

Guys,

Just to let you know, I've got an Qstarz BT-1000, this used to work flawlessly with my Herald and Garmin XT.
Now using the Touch HD it also lags behind.
I was reading your tries to get GPS working better, also tried some of the tweaks and it worked a little bit better.
But seeing the things on the Qstarz, isn't it likely the problem is somewhere else than in the GPS stack?

Rom 1.14.404.3 (22273) NLD 10/25/08
Radio 1.09.25.14
Prot 52.494.25.26U

Just my 2c, Cheers,

Edward

[X2D]

@X2D: the cab primarily plays with MaxBufferSize in Multiplexer, PollInterval in GPSOneDevice, and SleepOnNoData in Drivers. It also turns off logging. They are the settings I have on my Fuze.



Will give your first cab a shot then hard reset and try the second cab, when I get home the only thing I'm going to mess with is GPS and continue to watch this thread.

Da_G's stripped ROM
- Hard Reset upon first Flash, Update QuickGPS, Install/Run GPS Test to initiate the cold start.
- Run various Nav apps for a test drive (I notice myself moving as soon as I leave my driveway but do you guys know if it takes any amount of time for the GPS to essentially "sync" with my position? I wouldn't think so as it should have already done that with the lockon with GPS Test and stay locked on when running the Nav software)

-- Based on the above test results if it's still failing I will do another Hard Reset, Install your first cab, Update QuickGPS, Install/Run GPS Test which should already be warm
-- Again run various Nav apps to test position delay

--- Based on the last test results if it's still failing I will do another Hard Reset, Install your SECOND cab, Update QuickGPS, Install/Run GPS Test which should already be warm
--- Run various Nav apps to test position delay


Two questions, do the first and second cabs have a massive difference?
Second question is should I close GPS Test or continue to keep it running while I'm in the car.

I'll think of more questions and try to do as much testing as I can, I'm kind of a newb at all of this and I'm learning as I go but let me know if I can be more effecient and I will do whatever I can to post results, and do specific testing.


Edit:

Guys,

Just to let you know, I've got an Qstarz BT-1000, this used to work flawlessly with my Herald and Garmin XT.
Now using the Touch HD it also lags behind.

This is one thing I've wanted to try but I don't have an external GPS receiver. So you are STILL receiving the position lag where it shows you a few meters behind where you really are?

Does it delay a couple seconds when going from any MPH to 0?

[Sleuth255]

@aevh: Ok, that's a Bluetooth GPS receiver. How did you set it up? The traditional way is to add to paired devices then create a new outbound serial port and attach to the BT GPS device you just created. Use the BT manager for this. Next, set your navigation software to use the port you just created. The GPSID doesn't generally get involved here.

Are you sure that you aren't unknowingly still using the internal GPS ;)

[Sleuth255]

Two questions, do the first and second cabs have a massive difference?
Second question is should I close GPS Test or continue to keep it running while I'm in the car.


The second cab is identical to the first except that I've pulled all device specific references (such as the name of the GPSID driver dll). Don't install the first cab because it will break your GPSID unless Da_G named the new dll gpsid_htc.dll.

You must close GPS test once you have obtained a fix. Make sure the navigation software is all that's using the GPS when you go driving.

The rest of you with position lag are doing this right? If you have two programs accessing the GPS stream then it might contribute to position lag.

[hopper13]

haha... not sure I'm up for trying to security unlock and flash the radio yet, I'm an "end user" type. ;-) That said, I may have poked and prodded enough to require the dreaded hard reset and reinstallation of tons of stuff soon anyway... things are less stable than the once were.

Here is today's data. All results are agps-off and obtained using TomTom 7.9 after a soft reset, cab installation and power off/on. Skys are clear, time to lock is with the phone on the dash of my Land Cruiser (too icy for the Mini today) facing south giving the phone a clear view of approx 180 degrees azimuth and 90 degrees of vertical inclination.

Time to initial lock: 8:32
Time to subsequent locks: 0:17
Speed lag @ 22 mph: 5 seconds
Speed lag @ 44 mph: 8 seconds
Position lag: <1 sec to 1.8 seconds tested with speeds up to 45 mph.

Conclusions. The position lag does seem to have improved with this cab and imho is now quite acceptable. It should be noted that position lag appears to be somewhat variable, better when speed is decreasing and worse under acceleration. I have attempted to take all observations when the speed has been stabilized for at least 10 seconds prior.

Thanks to Sleuth and others for their hard work!!!!!!

Just my .02... your lag may vary....

hopper

[petard]

Could someone with some navigation software on their laptop conduct a test?
I think it would be good to get one of those programs that outputs the GPS data to a computer, and use that on a laptop to see if that also lags.

[X2D]

The second cab is identical to the first except that I've pulled all device specific references (such as the name of the GPSID driver dll). Don't install the first cab because it will break your GPSID unless Da_G named the new dll gpsid_htc.dll.

You must close GPS test once you have obtained a fix. Make sure the navigation software is all that's using the GPS when you go driving.

The rest of you with position lag are doing this right? If you have two programs accessing the GPS stream then it might contribute to position lag.


Will give the stripped rom + second cab a shot and report back, might make a trip home to install this ROM lol, really anxious to get to some real work instead of ya know, my job. :)

[Sleuth255]

@hopper13: Thanks for testing :) "ylmv" lol...

Security unlocking isn't for the faint of heart. It requires flashing an "intermediate" radio first and can be a real nail-biter.
I'll re-run my tests with acceleration involved in the position lag determination. I need to find a nice straight road with lots of side streets for the side-by-side tilt/fuze test. Based on how TomTom uses $GPVTG data (with known speed lag), I can see position lag increasing under acceleration then "catching up" once cruising speed has been obtained. I would imagine that this artifact affects my Tilt as well however since it also exhibits speed lag.

edit:

Actually, my video already has that. Granted it wasn't 0-60 in 5 seconds acceleration testing but it did show that my Tilt and my Fuze tracked pretty much identically as I accelerated normally up to 45-50mph. This means that if position lag increases during acceleration then both devices show it on the map about the same.

[btprice2001]

I'm testing Da_G's rom, I just flashed it and I have Visualgpsce running right now. When I run Tomtom will I have to set it to other nmea/comm 4 instead of built-in gps?

[Sleuth255]

I set it to "Built-in GPS". If you get position lag, try tossing in GPSIDregfix.cab as well.

[btprice2001]

I tried both built in and other NMEA and Tomtom says "No GPS device!" Google MAPS and Visualgpsce worked just fine. Any thoughts?

[maw]

I think you have to hard reset your device. I had that problem as well, seems that something got deleted in your registry that shouldn't have been deleted.... so hard reset, restore your backup and try something else. I hope you made a backup....

[aevh]

Hi Sleuth255,

@aevh: Ok, that's a Bluetooth GPS receiver. How did you set it up? The traditional way is to add to paired devices then create a new outbound serial port and attach to the BT GPS device you just created. Use the BT manager for this. Next, set your navigation software to use the port you just created. The GPSID doesn't generally get involved here.

Are you sure that you aren't unknowingly still using the internal GPS ;)

At first I did, than after some fiddling it got instead of my default Com4 (4800) none to BT GPS receiver inside the Quarz. The pairing was no issue, I set up a Com 5 for it outward, but GXT didn't find it, so I told it to get the BT GPS instead, when I switched off the Quarz I got the message from GXT that it was missing it's GPS fix, so technically I did get it right (I suppose).

What I saw was that the external GPS mouse was just as bad as the internal one, but maybe somebody can check their old GPS mouses with this kind of setup. I'll however do some testing tomorrow.

Cheers and thanks,

Edward

[Sleuth255]

I tried both built in and other NMEA and Tomtom says "No GPS device!" Google MAPS and Visualgpsce worked just fine. Any thoughts?

That could be normal. Mine does it too when I first load up TomTom. Wait a few seconds (usually about 10 and it should switch to searching).

[btprice2001]

Nope, I waited minutes and it stayed on No GPS device. I just hard reset per maw's advise. Let's see what that does.

[Da_G]

Curious, maybe gpsid_htc.dll mangles the nmea sentences in some way to make them 'more compatible'

I see the position lag in google maps, so that's what i'll be using to test when I go to starbucks in a minute..

[noellenchris]

Nope, I waited minutes and it stayed on No GPS device. I just hard reset per maw's advise. Let's see what that does.

I noticed after I clear the GPS or power off, the GPS does a cold start and can take a long time (up to 10min) to find some sats. It only happens sometimes. Anyhow I'm doing the same thing....waiting.

[btprice2001]

Nope, Tomtom still stayed on "No GPS device!" after hard reset. Shouldn't it switch from No GPS to waiting for valid signal or something? I even tried Tomtom 6 too. I guess I will try to test with Google Maps then.

[Sleuth255]

Can you turn on logging with advanced config then post some logs? I can look at the nmea sentences to see what's up.

also try starting with gps test then leave it running & run TomTom. If tomtom sees the gps, close gps test.

[btprice2001]

Sure, I'll post some logs. I also tried leaving Visualgpsce running with Tomtom and no difference.

[Ninja1]

The two of you who uploaded GPSID keys all have a few things in common. I'm going to make a cab for you to try.

Edit: Here's the cab. Let me know if it affects your position lag.

OK, added this and did a couple of tests. Time to first fix about 90 seconds--last download from QuickGPS this am (auto download at 10am). Went for a quick walk with the dog and tracking was very accurate in Google Maps and Live Search. Will go for a quick drive during half time.

To limit variables I am still running DaG 2.06. Will try the MS GPSID build later if needed.

[Da_G]

Testing GPSID_HTC is great, but wouldn't testing the MS GPSID eliminate GPSID as a source of lag alltogether (or prove that it IS the source of lag)

the MS GPSID is known good (it's used in many production builds that do not have this lag)

So if MS GPSID has the same lag, tweaking all these params isn't gonna do a lick of good since the source is more low level than that..

but if MS GPSID does not have the same lag, it's worthy to persue tweaking gpsid_htc.dll to try and equal it.. or just use it instead :)

[Sleuth255]

True, but I don't have position lag. My proof is on post 400. Wouldn't that also say that GPSID isn't the issue? I don't see why we can't pursue both avenues. Any data is good data.

[petard]

I installed GpsGate on my FUZE and my laptop. I have MapPoint 2009 on my laptop, so I passed the GPS data on to that. I am seeing the exact same position (and therefore lag) on my laptop and on my FUZE (it was running TomTom at the same time the laptop was tracking in MapPoint 2009).
I think this proves that it is for sure not a graphics or navigation software issue. The problem lies before the outputted GPS data.

[Da_G]

Of course, I only made the post because he said "if needed" so I wanted to make sure it's understood that it is needed :)

Because if the problem still occurs with MS GPSID, we're barking up the wrong tree.. need to look at gpsdriver.dll or the radio rom itself

Personally I think it's some averaging algorithm that needs to be toggled off with a bit in the radio configuration NV.. i've been trying to get into that for some time now but I haven't had luck.. can't just access it through MFG bootloader like we could on the kaiser..

Edit: Here's a build with all HTC GPS stuff stripped except QuickGPS.. it does the epheremis injection in the background before you even connect to the gps, so I don't see why it would cause the lag.. so at least the initial fix is nice and fast still :P
http://rapidshare.com/files/182179124/2.07-pre-agps-stripped-with-quickgps.rar

[btprice2001]

I enabled logging in Advanced Config and let GPS Test run for a while but the text file is blank. In Advanced Config I set Log File: \GPSLogFile.txt and Max. log file size: 32768. Do I need to do anything else?

[Sleuth255]

IMHO, there's three issues with Raphael GPS and they are all unrelated:

Initial Fix times
Position Lag
Speed Lag


I think that point 1 is solved. It's radio related. To see what I mean you need to be security unlocked because you need to observe the way an initial fix happens using the PRN slots panel in VisualGPSce with a Raphael then with a Blackstone radio. Prior to blackstone, there are always 16 slots shown (satellites visable) with PRN's on 8. These are random PRNs though and they switch around until one gets a signal. Then when signal is above 28 for a given time, the display readjusts to show actual birds visable. This is a cold-start sequence. It looks like ephemeris data isn't being properly injected. The Blackstone radio shows a whole different sequence.

Position lag looks like a GPSID FIFO issue which is why adjusting maxbuffersize in multiplexer appears to affect it. I'm looking forward to reviewing the logs from the MS_GPSID because I was under the impression that nmea strings themselves are generated by the radio and accessed through the RIL by the gpsdriver. This is supported by the fact that the sentence sequence changed when I loaded the blackstone radio.

Speed lag is generated by the chip and expressed in $GPVTG. It doesn't have much effect on navigation so I'm not very focused on it.

[Sleuth255]

I enabled logging in Advanced Config and let GPS Test run for a while but the text file is blank. In Advanced Config I set Log File: \GPSLogFile.txt and Max. log file size: 32768. Do I need to do anything else?

Make sure logging is enabled then see if data appears. Once you have a fix let it run for a bit then shutdown. I'd like to see actual fix data. thx!

edit: hang on while I check msdn to see where the enable setting is...
edit2: hmmm... just having the name should enable logging. Did you power off then back on?
check the key HLKM\System\CurrentControlSet\GPS Intermediate Driver
the filename should be in CurrentLogFile

[noellenchris]

Position lag looks like a GPSID FIFO issue which is why adjusting maxbuffersize in multiplexer appears to affect it. I'm looking forward to reviewing the logs from the MS_GPSID because I was under the impression that nmea strings themselves are generated by the radio and accessed through the RIL by the gpsdriver. This is supported by the fact that the sentence sequence changed when I loaded the blackstone radio.



I had some trouble getting a fix after setting maxbuffersize = 128. I will resume after this Giants vs Eagles game...lol Originally changing this setting to 256 from 512 halfed my pos lag. Maybe coincidence? Furthur testing should reveal.

[btprice2001]

Figured it out, I didn't have anything for Old log file. Here's the log file. I'm more than happy to test anything specifically, so let me know if there's anything you'd like me to try.

[Ninja1]

Of course, I only made the post because he said "if needed" so I wanted to make sure it's understood that it is needed :)

Because if the problem still occurs with MS GPSID, we're barking up the wrong tree.. need to look at gpsdriver.dll or the radio rom itself

Personally I think it's some averaging algorithm that needs to be toggled off with a bit in the radio configuration NV.. i've been trying to get into that for some time now but I haven't had luck.. can't just access it through MFG bootloader like we could on the kaiser..

Edit: Here's a build with all HTC GPS stuff stripped except QuickGPS.. it does the epheremis injection in the background before you even connect to the gps, so I don't see why it would cause the lag.. so at least the initial fix is nice and fast still :P
http://rapidshare.com/files/182179124/2.07-pre-agps-stripped-with-quickgps.rar

OK, OK! I know that it is needed, but like Sleuth I think that any data is good data to have. Just want to avoid flashing HTC/MS builds back and forth to test this and that. Would rather stay on 1 buils and exhaust it before moving on.

Half time, score is 10 to 8. Time to go for a drive.

[maw]

Sleuth, noellenchris en Da_G

Got a question since you both 3 are using the BS radio. Do you all have this when using GPS that its going off and on all the time?? I find it weird, I haven't had this problem with my previous radio, can it damage the GPS chip or not?

Let me know if you all have this problem.

[Da_G]

What do you mean going on and off all the time?

If you mean it reports you a fix, then doesn't give you any data for a period of time, then says 0 sats, gets a fix again, and then gives you no data.. in a loop, that's the behavior i got when i turned the maxbuffersize too low..

[maw]

Yes, thats exactly what I mean!

My BufferSizes are 512, what do you have? Also under multiplexer is 512

[Da_G]

I don't recall exactly what the value was to cause that behavior, it's been some time since I tweaked them..

But raise it higher and it should go away :)

[maw]

Can you check in your registry please?

I changed MaxBufferSize under Multiplexer first to 1024, no change, then to 2048, still no change.

Must I change the other buffers as well?

[Da_G]

I don't have any values at all for buffering in my registry, only poll interval set to 3e8 (1000)

I'm using the GPS clean build I posted earlier, with CE OS 5.2.21014.1.6.0..

[Sleuth255]

Figured it out, I didn't have anything for Old log file. Here's the log file. I'm more than happy to test anything specifically, so let me know if there's anything you'd like me to try.

Way cool stuff btprice2001 :)

You log contains no $GPVTG! :eek: That means that HTC GPSID is deriving speed and heading info from position data. TomTom can't navigate without $GPVTG so issues no gps device found. No wonder there's speed lag...

Heh... the NaN error is all HTC too :rolleyes:

[Sleuth255]

Sleuth, noellenchris en Da_G

Got a question since you both 3 are using the BS radio. Do you all have this when using GPS that its going off and on all the time?? I find it weird, I haven't had this problem with my previous radio, can it damage the GPS chip or not?

Let me know if you all have this problem.

Not all the time. When I have good signal the lock sticks. If I have weak signal (like inside a house) then it can bounce in and out. Fire up visualgpsce and watch the sequence inside and you'll see what I mean.

[mindfrost82]

Mine will get a fix and lose it after a second, but it usually only does this for about 15-30 seconds, then it'll lock on and keep it.

[Ninja1]

OK, I still have position and speed lag, but it is better than with the stock 2.06 settings. Another odd thing...both Google Maps and Live search show my position as consistently about 30 feet or so west of my actual position.

iGuidance accurately shows my position when at rest. Speed lag is consistent at about 3 seconds. Positional lag runs about 1 second at 30mph, and increases about 1/2 second per 10mph.

[mindfrost82]

OK, I still have position and speed lag, but it is better than with the stock 2.06 settings. Another odd thing...both Google Maps and Live search show my position as consistently about 30 feet or so west of my actual position.

iGuidance accurately shows my position when at rest. Speed lag is consistent at about 3 seconds. Positional lag runs about 1 second at 30mph, and increases about 1/2 second per 10mph.
That sounds like what I have with the Da_G 2.06 ROM already...which is usable, but still off compared to my Tilt.

My lag has never been very bad, but it is noticable.

[Sleuth255]

@ninja1: could you post a reg of your HKLM\System\CurrentControlSet\GPS Intermediate Drivers Key and all subkeys/values plz?

Just load up dotfred go to the key, tap/hold and choose export.

[btprice2001]

Way cool stuff btprice2001 :)

You log contains no $GPVTG! :eek: That means that HTC GPSID is deriving speed and heading info from position data. TomTom can't navigate without $GPVTG so issues no gps device found. No wonder there's speed lag...

Heh... the NaN error is all HTC too :rolleyes:

Not really sure what any of that means but glad I could help:) I could use Tomtom just fine with Da_G's regular roms btw. It's only with the GPS-stripped version that I get "No GPS device!"

[Da_G]

TomTom is probably looking for the GPVTG sentence and not finding it, so it's returning that error.. google maps can operate fine without that sentence though. How odd that it's not coming through.. the gps should be generating that..

[noellenchris]

Sleuth, noellenchris en Da_G

Got a question since you both 3 are using the BS radio. Do you all have this when using GPS that its going off and on all the time?? I find it weird, I haven't had this problem with my previous radio, can it damage the GPS chip or not?

Let me know if you all have this problem.

that happened earlier when I set the Max setting to 128.. I switched back to 256 and then I eventually went outside and after a few soft resets and pwr off's it found sats and started acting normal again. I'm going to try something between 128 and 256 next.

[Sleuth255]

TomTom is probably looking for the GPVTG sentence and not finding it, so it's returning that error.. google maps can operate fine without that sentence though. How odd that it's not coming through.. the gps should be generating that..

I would have expected that too. Also: no "GPVTG" string found in gpsid_htc.dll. very strange...

edit: I guess that if gpsid_htc.dll isn't generating it then gpsid_ms.dll might be supressing it from the raw data stream for some reason.

[Ninja1]

As requested.

[Sleuth255]

Thx, no surprises there. How do you have iGuidance set up? COM4: and at what speed?

edit: crap.... whatever happened to Eli Manning today????

[Jouke74]

For all of you which have the GPS going on and off check the value of SleepOnNoData. If this value is lower then your polling interval, it means that your GPS will be turning on and off constantly. This is probably not good for the GPS chip, and also not for obtaining a fix.

Eg. SleepOnNodata = 100 and PollingInterval = 300 means that your GPS will be turned off two times during one PollingInterval... at 100 and 200 milliseconds it does not get data, so it goes to sleep, then on the third it gets data from the satellite polling.

Put the SleepOnNoData back to 1 second, which is 1000. For the Lag and Speed it makes little to no difference, but it does for the first fix (I measured).

I posted this already earlier, but people seem to have ignored my post.

Comparing my HTC (Google maps) with my TomTom HD one it shows nearly identical readings (in speed and lag) in my car having around 9 sattelites driving 80 Km / H with this registry settings:

- HKLM\SYSTEM\CurrentControlSet\GPS Intermediate Driver\Drivers\GpsOneDevice\PollInterval = 250
- HKLM\SYSTEM\CurrentControlSet\GPS Intermediate Driver\Drivers\InputBufferSize = 128
- HKLM\SYSTEM\CurrentControlSet\GPS Intermediate Driver\Drivers\OutputBufferSize = 128
- HKLM\SYSTEM\CurrentControlSet\GPS Intermediate Driver\Drivers\SleepOnNoData = 1000
- HKLM\SYSTEM\CurrentControlSet\GPS Intermediate Driver\Multiplexer\MaxBufferSize = 256

COM4:
Baud = 4800

ROM all original:
Radio 1.02.25.19
Rom 1.90.478.2 Dutch
Protocol 52.33.25.17U

PS. I don't have TT7 for my HTC and I won't install an illegal version.

[Sleuth255]

Nice post. Forgot about that little ideosyncracy. Funny that it didn't crop up until I loaded the Blackstone radio though. Testing now...

edit: well, the fix bounced once then locked while I was inside the house.. That's more or less what I was seeing before.
edit2: yep confirmed with a hot fix. Lock still bounces in an out a bit using visualGPSce then locks after a few seconds.

[Da_G]

Ok, I just got done doing a bit of driving with Google Maps + MS GPSID, and the position lag is most definitely still there. I did 3 soft resets while I was out and I noticed differing behavior each time (one time it was off by about 30 ft as someone mentioned before, ive also seen that before)

So i'm going to say that all the tweaks we are doing are probably just creating a placebo effect more than anything.. unless gpsid_htc.dll is passing the values farther up the tree. Don't think so, but rather than guess, i'm starting to poke my nose into gpsdriver.dll..

I'm thinking either gpsdriver.dll or the radio rom, at this point.

[Ninja1]

Thx, no surprises there. How do you have iGuidance set up? COM4: and at what speed?

edit: crap.... whatever happened to Eli Manning today????

Com 4 @ 9600.

As for Eli, WTF? It was like watching a slow death...

[mindfrost82]

Don't think so, but rather than guess, i'm starting to poke my nose into gpsdriver.dll..

I'm thinking either gpsdriver.dll or the radio rom, at this point.

No way we could pull some of those dll files from the Tilt could we?

I have no position lag at all on my Tilt.

[Da_G]

It's worth a shot, i suppose. Find me a link to the latest tilt official ROM and i'll yoink the gps related stuff and see what's compatible.

[mindfrost82]

It's worth a shot, i suppose. Find me a link to the latest tilt official ROM and i'll yoink the gps related stuff and see what's compatible.
Here ya go: http://www.htc.com/us/SupportDownload.aspx?p_id=67&cat=2&dl_id=94

Just do the two dropdown boxes at the bottom and click on the button.

[btprice2001]

Here is the latest Tytn II as well

http://member.america.htc.com/download/RomCode/HTC%20TyTN%20II%5CRUU_Kaiser_HTC_WWE_3.28.405.0_ra dio_sign_25.83.40.02_1.65.17.56_Ship.exe

[Da_G]

Right-o! It's third on my list after fixing those 2 broken cabs in my ROM thread.. i might be out for the rest of the day so i'll probably get to it tomorrow. :)

[Sleuth255]

Com 4 @ 9600.

As for Eli, WTF? It was like watching a slow death...

Probably won't affect anything, but I'd try bumping the speed to 115200.

re Eli: Mob must've told him to throw the game....

[X2D]

Flashing Da_G's rom then going for some runs around and probably to the store to get food. Will edit this post or reply with results with absolutely no tweaks just GPS Test preinitialize after quick GPS downloads then after the lock on close GPS Test and try Garmin on the way there then iGO on the way back, if I see no change in position lag im going to install the second cab listed for the GPS tweaks. All while DRIVING! This is safe and effective ;)

[Sleuth255]

It'll be interesting to see if those navigation pgms see the gps. The test rom doesn't issue $GPVTG statements for some reason. TomTom won't run without that sentence. Let us know!

[Ninja1]

Probably won't affect anything, but I'd try bumping the speed to 115200.

re Eli: Mob must've told him to throw the game....

OK, bumped to 115200. This changed the warm TTF to 3 seconds while in the house. Not sure about lags as I am watching SD and Pitts play...can't let anything interrupt the Playoffs!

[Ninja1]

Oh Oh. The wife and kids are home...time to be a Dad and not a tester. See you all 'morrow noc.

[X2D]

GPS Log attached is from Beeline GPS, one of the only two programs that worked with this setup (did not try google), Garmin XT allowed me to lock on after a cold start as did Beeline GPS. Unfortunately iGO did not work. I did not have Tom Tom installed on my card so I did not test that and this is without the fix cab. Going to test that in a little bit.


Edit:

Just went on a further drive, installed the GPSIDRegFix and turned device off for 15-20 sec and turned it back on, 30 min later and it would not get a fix. GPS Test would come really close (or what it looks like) to getting a fix or get a fix for maybe 1-2 seconds then drop.

Uninstalled, turned device off etc, within 6 seconds got a fix drove home with the nice Garmin lady that told me the best path to take. Sadly I would pass an intersection or an overpass or make a turn and still experience a 2-3 second delay no matter what speed.

[noellenchris]

On my test today, TT7 was on the money. Hard to count lag. 1 sec or less passing roads. I shutdown TT7 and loaded up my Garmin XT. I had consistent 3-4 sec lag. I will test Garmin again in AM on the way to work using the Intermediate driver instead of the Com4 option. Not sure if it makes a difference, but I thought I read somewhere that using the intermediate driver worked for someone. We'll see. I'm still using the settings below.

HKLM\System\CurrentControlSet\GPS Intermediate Driver\Multiplexer

MaxBufferSize = 256

HKLM\System\CurrentControlSet\GPS Intermediate Driver\Drivers

InputBufferSize = 256
OutputBufferSize = 128
SleepOnNoData = 1000

HKLM\System\CurrentControlSet\GPS Intermediate Driver\Drivers\GPSOneDevice

PollInterval = 250

My Log files are set to 0.

My Performance settings are:

File sys cache = enabled
File sys cache size = 8mb
File sys filter cache = 131072
Glyph cache = 128KB

[Sleuth255]

Well, it looks like you're reproducing my TomTom results now. Now there are (at least) two. :)

It also looks like TomTom and other navigators differ with regards to position lag. I was wondering about this.

Is TomTom set to use the Built-in GPS device or COM4:? If its COM4:, what speed did you set?

[noellenchris]

Well, it looks like you're reproducing my TomTom results now. Now there are (at least) two. :)

It also looks like TomTom and other navigators differ with regards to position lag. I was wondering about this.

Is TomTom set to use the Built-in GPS device or COM4:? If its COM4:, what speed did you set?

It is using the built in GPS... is that the same as the garmin Intermediate driver? I will see 2morrow. Hopefully no lag (thinking out loud) I own Garmin XT and the 2009 maps for my wife's Garmin. I don't own TT7 and I will only be able to nav in my town with free map...lol

gnite..

[maw]

Just wanted to say that the BS radio is awesome! Woke up this morning and my battery level didn't budge one bit! COOL!

Now to test the GPS....

[maw]

Just tested the GPS, still a lot of lag! I'm going to try the settings from noellenchris and see if that reduces the lag. I still have position lag....

[land2634]

It is using the built in GPS... is that the same as the garmin Intermediate driver? I will see 2morrow. Hopefully no lag (thinking out loud) I own Garmin XT and the 2009 maps for my wife's Garmin. I don't own TT7 and I will only be able to nav in my town with free map...lol

gnite..

Just wanted to state that on several different devices, Garmin always had HORRIBLE lag that led me to believe it was software related. Other people I know noticed the same. If I had to estimate, the lag was about 8-10 seconds at least, regardless of device. The position would only update once every 2-2.5 seconds.

[maw]

I tested the GPS with noellenchris's settings and no shit, almost 0 lag! WOW! It trails about a second position wise and full stop from 50 to 0 in 4-5 seconds!

I really got to say, this BS radio is the bomb!
- better sound reception/quality
- better GPS accuracy
- better battery!

Damn... why didn't they make this radio for the Touch Pro as well?! It would save a lot of people headaches and don't even mention the hours! I was wondering, can't Da_G take apart this radio and extract the drivers necessary and put them in a Raph radio?? But there are so many good things in this radio... Let's see what HTC comes out with...

[Brendo]

Well I've been using stock rom and radio, so now I'm gonna try give a new radio a crack since my current one seems not to even get detected (always shows 'no gps device')

[mindfrost82]

I'm still having the same amount of lag, even with the new reg settings. Maybe mine was never as bad as some of your's, but its still there, about a second or so, which, as I mentioned before, is still usable.

No registry setting I make seems to make any noticable difference at all with my lag.

One thing I noticed with my speed lag is that the cursor usually stops within a second or so, but the speed display is still 3-4 seconds behind that. So even though the cursor stops almost when I do, it'll still show I'm going 5-10mph for a few seconds.

My position lag has almost always been 1-2 seconds, no matter what registry changes I've made.

I'm going to wait for Da_G and hopefully he can get some of the DLL's from the Tilt(Kaiser) to work and we can test that theory.

Sleuth--If we would stop getting snow I would make a video like you did, but its too dangerous with snow on the road since I would have to hold my camcorder in one hand lol.

[s.bonito]

On my test today, TT7 was on the money. Hard to count lag. 1 sec or less passing roads. I shutdown TT7 and loaded up my Garmin XT. I had consistent 3-4 sec lag. I will test Garmin again in AM on the way to work using the Intermediate driver instead of the Com4 option. Not sure if it makes a difference, but I thought I read somewhere that using the intermediate driver worked for someone. We'll see. I'm still using the settings below.

HKLM\System\CurrentControlSet\GPS Intermediate Driver\Multiplexer

MaxBufferSize = 256

HKLM\System\CurrentControlSet\GPS Intermediate Driver\Drivers

InputBufferSize = 256
OutputBufferSize = 128
SleepOnNoData = 1000

HKLM\System\CurrentControlSet\GPS Intermediate Driver\Drivers\GPSOneDevice

PollInterval = 250

My Log files are set to 0.

My Performance settings are:

File sys cache = enabled
File sys cache size = 8mb
File sys filter cache = 131072
Glyph cache = 128KB

THANX a lot, that works fantastic... after a few seconds it was synchronized. I'm using Garmin Mobile XT with [ROMeOS²] [GER] v1.70.3 (http://forum.xda-developers.com/showthread.php?t=430899).

[Sleuth255]

Since people have begun reporting success at curing this, I thought I'd post a summary:

I've been busy tracking reg settings and these are the ones that appear to be having the most impact:

MaxBufferSize in Multiplexer set to 256
PollInterval in GPSOneDevice set to 250
SleepOnNoData in Drivers set to 1000
CurrentLogFile and OldLogFile in GPS Intermediate Driver set to NULL

I have attached an updated Cab called GPSIDregfix.cab that sets these values.

ROM versions may impact this as well. It appears that the v2.0 or greater Diamond GPS drivers are involved with people who cure position lag

Radio versions may impact this as well. All posters who have cured position lag are running the 1.10.25.25 Blackstone radio. 1 Poster (me) has also achieved this with the 1.02.25.28 Radio. No posters have reported cured position lag using any other radio atm.

Also, it has been noted that different GPS Navigation software is impacted differently. Garmin software has been cited as being notorious for position lag. TomTom v7 can be achieve 0 position lag and is fully usable for navigation. The status of iGO is that settings/radio updates can reduce position lag to mostly 1 or less seconds. 1 poster has reported here and he considers it satisfactory and usable for navigation.

Edit: cab updated. It was pointed out that SleepOnNoData somehow was left at 100 with the original.
Edit2: now CurrentControlSet has a "0" in it lol. Fixed... what I deserve for changing cabs prior to 2nd cup of coffee :o

[maw]

I have been using only iGO8 version 8.3.2.64873 all the time. Like I said in my earlier postings TT7 hasn't been satisfactory to me. With the new BS radio and settings I have practically no position lag (just 1 maybe sometimes 2 seconds) and speed lag has a delay in the navigation when full stopping of about 4-5 seconds eventhough the position of the GPS has stopped on the screen.

BTW I see that noellenchris has his glyph cache set to 128 KB under performance in advanced config. I have mine set to 32 KB as advised, will changing it to 128 KB make it better??

[Sleuth255]

I've been looking at this myself. My 3.0 Sleuth Kaiser ROM doesn't have a gpsdriver.dll. Instead it has a gpsid.dll and a gpsid_qct.dll. The second dll is referred by the GPSID key and loaded in Drivers\Builtin. The first isn't loaded as a driver or service. Both are stored as modules in the ROM. Kaiser also uses a version 1 GPSID (ours is a version 2). Hopefully this GPSID will work with Raphael's RIL. I'm sure that either Da_G or I will have more on this itnf.

[Sleuth255]

@maw: I made the statement I did about iGO because you haven't declared it cured, merely acceptable for navigation. I would consider "cured" something equal to what I posted in the video. edit: (after re-reading the summary) I'll modify my post since you think its satisfactory.

Curing or reducing speed lag will be interesting. I need to find out what's going on with $GPVTG especially in light of the fact that either the HTC GPSID generates it in some as yet unknown way or the MS GPSID suppresses it for some as yet unknown reason.

[p50kombi]

Since people have begun reporting success at curing this, I thought I'd post a summary:

I've been busy tracking reg settings and these are the ones that appear to be having the most impact:

MaxBufferSize in Multiplexer set to 256
PollInterval in GPSOneDevice set to 250
SleepOnNoData in Drivers set to 1000
CurrentLogFile and OldLogFile in GPS Intermediate Driver set to NULL

I have attached an updated Cab called GPSIDregfix.cab that sets these values.

ROM versions may impact this as well. It appears that the v2.0 or greater Diamond GPS drivers are involved with people who cure position lag

Radio versions may impact this as well. All posters who have cured position lag are running the 1.10.25.25 Blackstone radio. 1 Poster (me) has also achieved this with the 1.02.25.28 Radio. No posters have reported cured position lag using any other radio atm.

Also, it has been noted that different GPS Navigation software is impacted differently. Garmin software has been cited as being notorious for position lag. TomTom v7 can be achieve 0 position lag. The status of iGO is undetermined at this point. Nobody has posted a satisfactory position lag outcome using iGO atm.

Curious, has anybody tried the new xperia radio rom? 1.10.25.18?
I've just got it loaded, but haven't had time yet to try it using TomTom.
I've added the reg edits manually, but to be honest, I always had very little lag.
So it's hard for me to compare.
Using pcfsabre's rom's..
I used the 1.10.25.28 radio, which performed pretty good before.
If anybody's interested I can post results tomorrow after driving home tonight with the .18 radio :)

ps. of the record, Sleuth, any progress in removing ezinput (or et9) but keeping fully functional hw keyboard?
I haven't had any luck yet, and tried loads of different things...

[p50kombi]

Since people have begun reporting success at curing this, I thought I'd post a summary:

I've been busy tracking reg settings and these are the ones that appear to be having the most impact:

MaxBufferSize in Multiplexer set to 256
PollInterval in GPSOneDevice set to 250
SleepOnNoData in Drivers set to 1000
CurrentLogFile and OldLogFile in GPS Intermediate Driver set to NULL

I have attached an updated Cab called GPSIDregfix.cab that sets these values.

ROM versions may impact this as well. It appears that the v2.0 or greater Diamond GPS drivers are involved with people who cure position lag

Radio versions may impact this as well. All posters who have cured position lag are running the 1.10.25.25 Blackstone radio. 1 Poster (me) has also achieved this with the 1.02.25.28 Radio. No posters have reported cured position lag using any other radio atm.

Also, it has been noted that different GPS Navigation software is impacted differently. Garmin software has been cited as being notorious for position lag. TomTom v7 can be achieve 0 position lag and is fully usable for navigation. The status of iGO is that settings/radio updates can reduce position lag to mostly 1 or less seconds. 1 poster has reported here and he considers it satisfactory and usable for navigation.

Sleuth, after looking into your CAB, I think your cab sets the SleepOnNoData to 100 instead of 1000..
Just thought I'd let you know :D

[Sleuth255]

The original cab did that. The one I just posted should have it reset to 1000. I'll modify the attachment in the original post.

edit: Crap this one has it too.... must not've saved the file b/f exiting notepad. Damn! uploading fix now...

SleepOnNoData hasn't been confirmed to change much however. I've been running at 100 forever. A recent change to 1000 appeared to make no difference. I included it because the GPSID references it and it just makes sense that SleepOnNoData should be greater than PollInterval.

[maw]

@sleuth: well if you put it that way, it ain't cured as yet... let me put it this way, it's getting better using this prescription, but it will need a proper vaccine (radio/rom) to completely cure it :) I kept to the medical terminology just for kicks LOL

[p50kombi]

The original cab did that. The one I just posted should have it reset to 1000. I'll modify the attachment in the original post.

edit: Crap this one has it too.... must not've saved the file b/f exiting notepad. Damn! uploading fix now...

SleepOnNoData hasn't been confirmed to change much however. I've been running at 100 forever. A recent change to 1000 appeared to make no difference. I included it because the GPSID references it and it just makes sense that SleepOnNoData should be greater than PollInterval.


Just thought I'd give you a heads up :)

[Sleuth255]

Slightly OT but I just noticed my battery on the second day after recharge is at 70%. :eek:

[Dyaus]

Been following this thread and doing some testing myself and figured I'd throw in my 2 cents.

In the last couple of days I switched to BS Radio 1.10.25.25 and PROven ROM 1.11. I also applied the registry settings noellenchris posted back a couple of pages. My position lag is now virtually zero. I do still have speed lag though. This is with TT 7.

I did notice that this setup seems to take longer to get a cold lock then my previous setup which was PROven ROM 1.09 and Radio v1.02.25.32. I used to be able to get a cold lock in a a couple minutes now it takes about 5 minutes to get a cold lock.

I usually fire up VisualGPSce until it gets a lock and then fire up TT or Google Maps and they'll get a lock pretty quickly (<20 seconds).

I don't really mind the speed lag, so having the positional lag working good makes me happy =).

I don't mind doing any tests to help out, so if anyone would like me to test anything, let me know.

[mindfrost82]

Sleuth: One of your cabs is messed up because now I have a key:

HKLM\SYSTEM\Cur0rentControlSet with some of the GPS settings in it lol.

[noellenchris]

BTW I see that noellenchris has his glyph cache set to 128 KB under performance in advanced config. I have mine set to 32 KB as advised, will changing it to 128 KB make it better??

I've played a lot with the glyph setting on my Hermes. Going to 32KB was definately noticeable update on the Hermes. So I guessed that the 128KB on the TP would help and it has plenty of RAM. Otherwise, I really haven't tested, I just assumed it would help.

Glad my settings seemed to help you. I did change the maxbuff setting to 128 today and really didn't notice a difference. Garmin Xt still is laggy with either the Intermediate driver or Com4 setting...hmmm

[maw]

Ok, I changed it to 128 KB.... I will check and see what kinda difference this makes. Maybe it will make my navigation work faster in the process hopefully getting rid of the last 1-2 seconds lag.... :)

Will report back....

[noellenchris]

THANX a lot, that works fantastic... after a few seconds it was synchronized. I'm using Garmin Mobile XT with [ROMeOS²] [GER] v1.70.3 (http://forum.xda-developers.com/showthread.php?t=430899).

Do you have position lag with your Garmin Mobile XT? I do...but not with TT7.

I'm looking into the program for the TouchHD that reduces the screen resoulution. I don't know how to get it to work on the TP. If I could reduce the resolution to 240x320 then Garmin may repaint the maps quicker....plus a lot of qvga apps would also work... Any ideas?

[X2D]

I am going to test the latest GPS fix cab you posted Sleuth with a different ROM again and then Da_G's again.

I am really trying to find a solution that isn't going to make me pay money to unlock my phone which honestly should be a free service but I can't complain, people have to make money for stuff so it's whatever, but if I happen to jack up this phone and send it back under warranty and reflash all of the original stuff then wouldn't that be wasted money? If I got a new phone would I have to re-unlock it?

Meh, will report back my findings with this latest GPS cab using the .28 Radio on T.i.R's ROM (reflashed last night out of boredom, I flash my phone at least once a day it seems like).

[s.bonito]

@noellenchris Not so far... But in the next days I'll drive a longer distance... Next Monday I can tell you more ;)

[noellenchris]

Does anyone know how to change the DPI? Or of software to change DPI, if we change that then we may be able to reduce screen resolution and speed up graphics....? Still digging into GArmin's lag.....:confused:

I found what I was looking for...... http://forum.xda-developers.com/showthread.php?t=327830

[Sleuth255]

Sleuth: One of your cabs is messed up because now I have a key:

HKLM\SYSTEM\Cur0rentControlSet with some of the GPS settings in it lol.


damn if it aint true lol.... WTF, can't type today...

fixed (if you can believe that)...

[Da_G]

I still think we need to be looking elsewhere than the GPSID settings for this problem.. microsoft GPSID is known not to have this issue regardless of settings, and it exhibits the same problems, to me that means gpsid_htc.dll is absolutely not the source of the issue(s)..

I'm looking at the tilt stuff now, will post back w/more info in a bit..

Wierd stuff on the GPVTG sentence btw. I can't find anywhere that it's generating it in the drivers.. really odd that msft gpsid isnt spitting it out.

[Da_G]

tilt uses taiserver.dll to handle gps commo rather than communicating directly to SMD: port which uses RIL, not too sure if it'll work, but it might.. id give it a 20% chance.. no time to port it now, on the way out the door :) I'll check it out more in depth in a few hours :)

[Black1982]

damn if it aint true lol.... WTF, can't type today...

fixed (if you can believe that)...


I installed that cab too and i have the same wrong registry key! Should
i just delete that key?

GPS still works btw.

[X2D]

Just installed it and it's fine for me, will test on the drive home with again Garmin, iGO, OCN8, TomTom, BeelineGPS and see what happens!

[DSF]

Been following this thread and doing some testing myself and figured I'd throw in my 2 cents.

In the last couple of days I switched to BS Radio 1.10.25.25 and PROven ROM 1.11. I also applied the registry settings noellenchris posted back a couple of pages. My position lag is now virtually zero. I do still have speed lag though. This is with TT 7.

I did notice that this setup seems to take longer to get a cold lock then my previous setup which was PROven ROM 1.09 and Radio v1.02.25.32. I used to be able to get a cold lock in a a couple minutes now it takes about 5 minutes to get a cold lock.

I usually fire up VisualGPSce until it gets a lock and then fire up TT or Google Maps and they'll get a lock pretty quickly (<20 seconds).

I don't really mind the speed lag, so having the positional lag working good makes me happy =).

I don't mind doing any tests to help out, so if anyone would like me to test anything, let me know.
How's on when changing directions? Does it 'hang' for a while then throw to actual position? (see Sleuth video).
If yes, try a route simulation. If in the simulation everything is ok.. then you still have positional lag

[Sleuth255]

I installed that cab too and i have the same wrong registry key! Should
i just delete that key?

GPS still works btw.

The bad key won't be used. Delete it if you like a clean registry I guess. You can also uninstall it then install the latest one I posted.

[btprice2001]

So I finally flashed the 1.10.25.25 Blackstone radio. First I tested with Tomtom using the default reg entries in Da_G's 2.06 rom (should've noted what they were). It might have just been a placebo effect, but I think position lag improved slightly. I feel like time to first fix improved too. I haven't noticed any difference (better or worse) in cell signal strength or battery life. Then I tried these settings:

HKLM\System\CurrentControlSet\GPS Intermediate Driver
CurrentLogFile: blank
MaxLogFileSize: 0
OldLogFile: blank
HKLM\System\CurrentControlSet\GPS Intermediate Driver\Drivers
InputBufferSize: 128
OutputBufferSize: 128
SleepOnNoData: 1000
HKLM\System\CurrentControlSet\GPS Intermediate Driver\Drivers
PollInterval: 250

Position lag was definitely improved if not completely gone! I seem to remember reading that PollInterval influeces battery life so I then tried increasing PollInterval to 500. Again, position lag definitely improved from before flashing this radio and using these settings, but maybe a little worse than PollInterval 250. PollInterval 500 seems to be very usable for navigation though and I'm going to test more with both values. Is it true that setting PollInterval higher will result in lower battery life?

[Edit] Leaving VisualGPSce running, my current is listed as -272 in TBattery at PollInterval 500 and -267 at PollInterval 250.

[Da_G]

Got my buddy with a kaiser to do some digging in his radio rom while I dig in the blackstone one, so far i'm not finding any GPS related configurations that are significantly different..

His tilt is configured to use a different AGPS server, and the radio rom doesn't have support for AGPS TLS..

but AT@setgps returns the same, as do a few other gps related commands..

Although I notice a setting the radio rom is pulling from sonewhere, htc_nmea_format, i think its stored in NV though.. so far the only nv related commands i've found is AT+HTCNV? which returns AT+HTCNV=1,12,8 on my ROM.. i broke that down and the params are param1=DTM support, 1 enabled 0 not, param2 = MCS support, valid values 8, 10, 12, and param3 = hsdpa category, valid values 6,8,12.. too bad it doesnt support 10 (hsdpa 14.4)

But i digress.. none of those particular settings are related.. digging further..

[mindfrost82]

Got my buddy with a kaiser to do some digging in his radio rom while I dig in the blackstone one, so far i'm not finding any GPS related configurations that are significantly different..

His tilt is configured to use a different AGPS server, and the radio rom doesn't have support for AGPS TLS..

but AT@setgps returns the same, as do a few other gps related commands..

Although I notice a setting the radio rom is pulling from sonewhere, htc_nmea_format, i think its stored in NV though.. so far the only nv related commands i've found is AT+HTCNV? which returns AT+HTCNV=1,12,8 on my ROM.. i broke that down and the params are param1=DTM support, 1 enabled 0 not, param2 = MCS support, valid values 8, 10, 12, and param3 = hsdpa category, valid values 6,8,12.. too bad it doesnt support 10 (hsdpa 14.4)

But i digress.. none of those particular settings are related.. digging further..
I don't think the aGPS matters. I don't have data with AT&T, and I didn't use aGPS on my Kaiser and I don't on my TP either and my Kaiser is still much more accurate when talking about lag.

[Da_G]

That's a 10-4, it doesn't matter to our lag problem, hence why I said nothing significantly different.. yet

I'm pretty convinced it's a simple NV bit that needs to be flipped.. for some kind of filtering.. especially since it seems the qualcomm equivilant of static navigation is on..

[Da_G]

$GPVTG is definitely in the radio rom, I wonder why it doesn't get sent by ms gpsid..

Check out these NV items, we need QPST so bad :P or use RegisterEditor from http://forum.xda-developers.com/showthread.php?t=451646 to dump nv and find address for these values.. so we can write new value :)

NV_AAGPS_DEVELOPMENT_TEST_CONTROL5_I
NV_CGPS_USE_FDIC_I
NV_GPS1_XTRA_ENABLED_I
NV_CGPS_DGPS_CORRECTIONS_ALLOWED_I
NV_CGPS_ALTITUDE_HOLD_MODE_I
NV_CGPS_REFERENCE_MODE_I
NV_CGPS_OPERATION_MODE_I
NV_AAGPS_IPC_DM_THREAD_MASK_I
NV_CGPS_ELEVATION_MASK_I
NV_CGPS_2D_PDOP_MASK_I
NV_CGPS_PDOP_MASK_I
NV_CGPS_MAXIMUM_DGPS_INTERVAL_I
NV_CGPS_STEERING_ON_STARTUP_I
NV_CGPS_MAX_ALMANAC_AGE_IN_WEEKS_I
NV_AAGPS_IPC_DM_THREAD_MASK_I

[kavi]

Been following this thread for a while and finally got myself a Fuze a few days ago.

Test results.

Radio 1.02.25.32
Tomom 7.910 (9185)
Map USA_Canada V725.1905 - No POI's enabled in MAP

1. Stock ATT ROM/Radio including bloatware + QGPS Updated

Cold Fix 1 - 2 min using gps test
Warm Fix - Under a min
TT7 drops signal every 10 -15 seconds not usable.


2. Stock ATT ROM/Radio + Bloatware + GPS Fixes (in the thread) - QGPS Updated

Cold fix - under a min
Warm Fix - under 30 seconds
TT7 works ok, 100 - 200 yards lag
Google maps works, 100 - 200 yards lag


3. Stock ATT ROM/Radio without running UC + QGPS Updated + AGPS ON (default settings)
Cold fix - under a min
Warm Fix - under 30 seconds
TT7 works, 100 - 200 yards lag
Google maps works, 100 - 200 yards lag

I was surprised to see the TP perform almost identical in scenario 2 & 3
Hope this will help the pros in finding the ultimate solution.

Side by comparison of my 100% Kaiser and TP (Scenario 3) Registry
http://i39.tinypic.com/2viovg1.jpg

[Ninja1]

OK, tested some stuff on the way home from work. Settings are as Noellenchris posted. Tested using iGuidance, which requires a com address and does not "see" the intermediate driver. Also used my Sirf III BT GPS puck as a check for the internal GPS. ROM is DaG 2.06.

Results:
Position in Google and MS Live is now about 30 feet east instead of west (WTF?).
Positional lag is better, but still proportional to vehicle speed and up to 100 feet off at 65mph.
Speed lag is about 3 to 4 seconds.
BT GPS shows the same results as the internal GPS.

So, maybe a graphics driver delay for some of this? iiGuidance does not use D3D, so is running on the unaccelerated 2D drivers. What do you think?

[kavi]

.

So, maybe a graphics driver delay for some of this? iiGuidance does not use D3D, so is running on the unaccelerated 2D drivers. What do you think?

Correct me if I am wrong but does Live and GMM use D3D? They both have lag issues even on mine.

[Ninja1]

AFAIK about the only thing that uses the D3D drivers are games.

[noellenchris]

So, maybe a graphics driver delay for some of this? iiGuidance does not use D3D, so is running on the unaccelerated 2D drivers. What do you think?

do they use OpenGL....there's a lot of talk on other posts about this. I'm still looking to reduce the lag w/Garmin.

[Sleuth255]

So I finally flashed the 1.10.25.25 Blackstone radio. First I tested with Tomtom using the default reg entries in Da_G's 2.06 rom (should've noted what they were). It might have just been a placebo effect, but I think position lag improved slightly. I feel like time to first fix improved too. I haven't noticed any difference (better or worse) in cell signal strength or battery life. Then I tried these settings:

HKLM\System\CurrentControlSet\GPS Intermediate Driver
CurrentLogFile: blank
MaxLogFileSize: 0
OldLogFile: blank
HKLM\System\CurrentControlSet\GPS Intermediate Driver\Drivers
InputBufferSize: 128
OutputBufferSize: 128
SleepOnNoData: 1000
HKLM\System\CurrentControlSet\GPS Intermediate Driver\Drivers
PollInterval: 250

Position lag was definitely improved if not completely gone! I seem to remember reading that PollInterval influeces battery life so I then tried increasing PollInterval to 500. Again, position lag definitely improved from before flashing this radio and using these settings, but maybe a little worse than PollInterval 250. PollInterval 500 seems to be very usable for navigation though and I'm going to test more with both values. Is it true that setting PollInterval higher will result in lower battery life?

[Edit] Leaving VisualGPSce running, my current is listed as -272 in TBattery at PollInterval 500 and -267 at PollInterval 250.

There are reports of reduced battery life that I've not been able to verify. What was your maxbuffersize setting in multiplexer?

[Sleuth255]

So, maybe a graphics driver delay for some of this? iiGuidance does not use D3D, so is running on the unaccelerated 2D drivers. What do you think?

It seems to be looking like graphics related lag contribution can be re-visited in light of the fact that different navigation products appear to show lag differently. A good test suggested earlier would be to use route simulation to see the effects of screen updates. Does it look ok when running simulated? Then maybe its GPS related...

[Sleuth255]

So, I actually observed position lag today. It happened right after I got a fix while I was driving. I was testing some changes that were purported to reduce speed lag. The issue cleared itself up automatically in less than a minute but not before I was able to closely observe it. Visually, when the problem cleared, the map "snapped" to the new position. By "snapped" I'm referring to the updates. If you look at my fuze/tilt comparison video, you can see that the movement of the fuze map is a bit "jerkier" than the Tilt. When my position lag corrected itself, the "jerk" moved the map ahead several hundred yards. I'm reviewing the reg changes now and will attempt to reproduce this again tomorrow.

[btprice2001]

There are reports of reduced battery life that I've not been able to verify. What was your maxbuffersize setting in multiplexer?

My maxbuffersize is 256.

Btw, did more testing of pollinterval at 250 on the way home. Position lag seems to be all but gone at lower speeds, but seems to reappear/get worse at higher speeds. My max speed during this drive was probably ~ 30mph. Still the lag was nowhere near as bad as before.

[btprice2001]

So, I actually observed position lag today. It happened right after I got a fix while I was driving. I was testing some changes that were purported to reduce speed lag. The issue cleared itself up automatically in less than a minute but not before I was able to closely observe it. Visually, when the problem cleared, the map "snapped" to the new position. By "snapped" I'm referring to the updates. If you look at my fuze/tilt comparison video, you can see that the movement of the fuze map is a bit "jerkier" than the Tilt. When my position lag corrected itself, the "jerk" moved the map ahead several hundred yards. I'm reviewing the reg changes now and will attempt to reproduce this again tomorrow.

I've meant to ask you about this before but always forget:rolleyes: Have you ever tried to use someone's settings that has position lag to try to reproduce it?

[Ninja1]

do they use OpenGL....there's a lot of talk on other posts about this. I'm still looking to reduce the lag w/Garmin.

Don't think so. I'll look around and try to figure it out.

[Sleuth255]

I've meant to ask you about this before but always forget:rolleyes: Have you ever tried to use someone's settings that has position lag to try to reproduce it?

You can bet on that happening now. :)

[Ninja1]

From whaat I can tell it looks like only the PC versions of iGuidance use D3D or OpenGL. Can't find anything that touts accelerated graphocs on the PPC.

Also, FYI for those who are interested...iNav is unbundling the PC and PPC versions of iGuidance. Thay are now seperate products...time to look for a new GPS program that will run on both platforms.

[mindfrost82]

My tests have been with iGo and Garmin 5.00.20 (the newest version, whatever that is).

One thing I've noticed with Garmin is that it constantly loses the GPS signal. If you stay on the main screen where the signal bars are you might notice this. While driving I'll go from 5 bars down to 3, then down to 1 red one, then it'll jump back up to 5. If I'm on the map I'll notice it getting stuck then jumping down the road, this is probably when it loses GPS.

I don't have this problem with iGo. It shows the signal strength on the map screen and its always full.

So this could be causing the lag with Garmin if you have the same problems...just no idea why its happening.

My latest test was with:
- PROven 1.11 ROM
- Blackstone 1.10.25.25 radio
- Sleuth's latest cab of GPS tweaks
- Applied to SD TuneUp cab from here (http://forum.xda-developers.com/showthread.php?t=446366)

I still had lag with Garmin, but I think its due to the problem I mentioned earlier.

My lag with iGo has probably been cut in half, now its down to about 1/2 second or so.

[noellenchris]

My tests have been with iGo and Garmin 5.00.20 (the newest version, whatever that is).

One thing I've noticed with Garmin is that it constantly loses the GPS signal. If you stay on the main screen where the signal bars are you might notice this. While driving I'll go from 5 bars down to 3, then down to 1 red one, then it'll jump back up to 5. If I'm on the map I'll notice it getting stuck then jumping down the road, this is probably when it loses GPS.

I don't have this problem with iGo. It shows the signal strength on the map screen and its always full.

So this could be causing the lag with Garmin if you have the same problems...just no idea why its happening.

My latest test was with:
- PROven 1.11 ROM
- Blackstone 1.10.25.25 radio
- Sleuth's latest cab of GPS tweaks
- Applied to SD TuneUp cab from here (http://forum.xda-developers.com/showthread.php?t=446366)

I still had lag with Garmin, but I think its due to the problem I mentioned earlier.

My lag with iGo has probably been cut in half, now its down to about 1/2 second or so.

My Garmin is like that too, I think it did that more when I lowered the polling to 250.

While driving my wife's car tonight, she has a stand alone Garmin unit. I noticed it too has about 2sec lag...lol. I never really noticed. So if I minus that 2 sec lag from my 3-4 sec lag on the TP it's about the same as with TomTom. The lag did seem less on my way home w/garmin on my TP. It's still on 128 for maxbuffersize. I don't really notice a difference between 128 & 256. I may change the polling to 500 again and see how that affects the lag. I think Garmin is updating (map refresh) at about once a sec. Kinda slow compared to iGuidance (my old favorite for routes), TT7 and I used to run a version of MioMap (loved the display and POI's but maps were older and no traffic) MioMap used to run the smoothest on my Hermes w/ext antenna.

[mindfrost82]

I think my Tilt had a little lag with Garmin too, I just don't like how it keeps dropping the connection, I think that makes it worse.

[noellenchris]

I think my Tilt had a little lag with Garmin too, I just don't like how it keeps dropping the connection, I think that makes it worse.

try changing the polling to 500. It did hold the signal good for me before, something I changed affected it. TT7 is steady with any of my settings. Garmin is finicky.... Wish they would update the refresh so it would be smooth like TT7/MioMap etc.

[p50kombi]

Hey guys.
I'm using the xperia radio, 1.10.25.18
Works a treat, fast lock and almost no lag.
It's still there, using tomtom7, but hardly noticable.
I think I can recognize sleuth's post that sometimes it starts lagging, and then "snaps" to place.
On highway the lag is nearly zero, driving 120 kmh.
When passing under bridge, Tomtom is almost spot on.
In the city ddriving between 30 and 50 kmh, it seems to be lagging a bit.
Using the settings from the cab (manually changed them though, didn't use the cab)
So in short, xperia radio performs great in Holland.
Great battery life as well and good quality.
Using Lantis OS 3.0 RC1 from pfcsabre..

If you want to know anything else, ask me...
Good luck in the search, I hope we'll be able to eliminate lag completely.

[Sleuth255]

Yep, you can see a bit of that in the video. That's the primary difference between my Raphael and my Tilt. That behavior would lead me to infer that the positional data from $GPGGA was just plain wrong for a while then suddenly corrected itself . I may write a test application to capture &parse fix data then throw up an alert dialog if adjacent $GPGGA sentence change vectors are outside of a given set of parameters. At the speed I was traveling, my position was off 40m or so during the time I was experiencing position lag.

[hopper13]

HKLM\System\CurrentControlSet\GPS Intermediate Driver
CurrentLogFile: blank
MaxLogFileSize: 0
OldLogFile: blank
HKLM\System\CurrentControlSet\GPS Intermediate Driver\Drivers
InputBufferSize: 256
OutputBufferSize: 256
SleepOnNoData: 1000
HKLM\System\CurrentControlSet\GPS Intermediate Driver\Drivers
PollInterval: 250

Sleuth,

After applying your cab, I modifed the inputbuffer, outputbuffer and sleeponnodata settings as above. Now when observing (in TomTom 7.9) the process of obtaining satellite reception, I see the signal strength bars as steady instead of going up and down and the time to lock after a reset this morning was under 2 minutes! Position latency seemed negligeble (sub 1 sec). Moving the input/output buffers above 128 seems to have helped a great deal.

Warning - this is emperically derived data with a sample size of ONE! Further testing is needed.

[mindfrost82]

My latest test was with:
- PROven 1.11 ROM
- Blackstone 1.10.25.25 radio
- Sleuth's latest cab of GPS tweaks
- Applied to SD TuneUp cab from here (http://forum.xda-developers.com/showthread.php?t=446366)
Alright, with this setup I'm pretty satisfied I think. I have yet to go on the highway to test though.

I did change the PollInterval to 500 as suggested, and now Garmin will hold the signal instead of constantly losing it.

Garmin is pretty even with my Tilt now. I wish Garmin would make an update to make the map flow instead of updating every second or so, but that's a Garmin thing, not the internal GPS.

iGo still has a little lag, but its less than 1 second for me most of the time.

PollInterval = 500
InputBufferSize = 256
OutputBufferSize = 128
SleepOnNoData = 1000
RetryWaitOnError = 3 (I don't think this one has been mentioned before, anyone know what it actually does?)
Mutliplexer MaxBufferSize = 256
MaxLogFileSize = 0
Current & Old Log Files are null

So at this point I'm satisfied, but I'm still willing to test further if any other suggestions come up.

[pandaboyy]

HKLM\System\CurrentControlSet\GPS Intermediate Driver
CurrentLogFile: blank
MaxLogFileSize: 0
OldLogFile: blank
HKLM\System\CurrentControlSet\GPS Intermediate Driver\Drivers
InputBufferSize: 256
OutputBufferSize: 256
SleepOnNoData: 1000
HKLM\System\CurrentControlSet\GPS Intermediate Driver\Drivers
PollInterval: 250

Sleuth,

After applying your cab, I modifed the inputbuffer, outputbuffer and sleeponnodata settings as above. Now when observing (in TomTom 7.9) the process of obtaining satellite reception, I see the signal strength bars as steady instead of going up and down and the time to lock after a reset this morning was under 2 minutes! Position latency seemed negligeble (sub 1 sec). Moving the input/output buffers above 128 seems to have helped a great deal.

Warning - this is emperically derived data with a sample size of ONE! Further testing is needed.

It's this the Cab (http://forum.xda-developers.com/showpost.php?p=3157513&postcount=623) that you use?

[nitroblu]

sorry if this is a repeat question, but based on what I interpreted from you post Sleuth, does this mean that if I am using a ROM with v.20 Diamond drivers, I do not have to apply the .cab you've posted to fix GPS lag? (btw thanks for all your works :))

[JJKameR]

Something damn wrong with my diamond...

GPS is working good with barely noticeable lag now on both iGo and Tom Tom.

But in Tom Tom instead of showing the satellites bars in the gps screen nothing is shown but position. Btw gps works good. Really better than before.

If it continues working like this I don't care... :D
I was just wondering if I am the only one that has this bug.

This settings in the reg:
HKLM\System\CurrentControlSet\GPS Intermediate Driver
CurrentLogFile: blank
MaxLogFileSize: 0
OldLogFile: blank
HKLM\System\CurrentControlSet\GPS Intermediate Driver\Drivers
InputBufferSize: 256
OutputBufferSize: 128
SleepOnNoData: 1000
HKLM\System\CurrentControlSet\GPS Intermediate Driver\Drivers
PollInterval: 250

[noellenchris]

Something damn wrong with my diamond...

GPS is working good with barely noticeable lag now on both iGo and Tom Tom.

But in Tom Tom instead of showing the satellites bars in the gps screen nothing is shown but position. Btw gps works good. Really better than before.

If it continues working like this I don't care... :D
I was just wondering if I am the only one that has this bug.

This settings in the reg:
HKLM\System\CurrentControlSet\GPS Intermediate Driver
CurrentLogFile: blank
MaxLogFileSize: 0
OldLogFile: blank
HKLM\System\CurrentControlSet\GPS Intermediate Driver\Drivers
InputBufferSize: 256
OutputBufferSize: 128
SleepOnNoData: 1000
HKLM\System\CurrentControlSet\GPS Intermediate Driver\Drivers
PollInterval: 250


That happened to me too..lol... haven't found out why though..I will keep playing with settings and see... If the darn thing would just work as advertized! lol

[Sleuth255]

sorry if this is a repeat question, but based on what I interpreted from you post Sleuth, does this mean that if I am using a ROM with v.20 Diamond drivers, I do not have to apply the .cab you've posted to fix GPS lag? (btw thanks for all your works :))

You do need to apply the cab as well.

[Da_G]

Did some more gpsdriver.dll debugging..

The radio rom is only spitting out a set of strings once per second. So when the pollinterval is cranked up higher, really we aren't getting a benefit..

Here's the debug output at 500: (2x per second)


17:09:49 [D:GPS] GPS: [ProcessPollData] --- g_hStateChangeEvent Received!
17:09:49 [D:GPS] GPS: [ProcessPollData] GPS data was received.
17:09:49 [D:GPS] GPS: [ProcessPollData] +++ Wait for g_hStateChangeEvent
17:09:50 [D:GPS] GPS: [ProcessPollData] --- g_hStateChangeEvent Received!
17:09:50 [D:GPS] GPS: [ProcessPollData] No GPS data was retrieved!
17:09:50 [D:GPS] GPS: [ProcessPollData] +++ Wait for g_hStateChangeEvent
17:09:50 [D:GPS] GPS: [ProcessPollData] --- g_hStateChangeEvent Received!
17:09:50 [D:GPS] GPS: [ProcessPollData] GPS data was received.
17:09:50 [D:GPS] GPS: [ProcessPollData] +++ Wait for g_hStateChangeEvent


You'll note that when polling twice per second, the gpsid ends up just hanging out on every other poll event, sucking up cycles for no particular reason because there's no data there to be had.

I haven't yet actually tested pollinterval 1000 and 100, but i can imagine at 1000 it'll be getting data every poll event, and at 100 there will be 9 poll events with no data and 1 with data.. but rather than conjecture, i'm off to test now :P

[Sleuth255]

Interesting. I'm sure seeing nmea data streaming at far faster than one sentence per second though. How to correlate this? Nmea data doesn't appear to be generated within gpsdriver (winhex only at this point, looking for strings). My xp with Hermes shows its generated by either the chip or the radio. Maybe gpsdata available is a buffer full of nmea strings, if so a single miss might cause 1 seconds worth of updates to be lost...

[noellenchris]

Did some more gpsdriver.dll debugging..

The radio rom is only spitting out a set of strings once per second. So when the pollinterval is cranked up higher, really we aren't getting a benefit..

Here's the debug output at 500: (2x per second)


17:09:49 [D:GPS] GPS: [ProcessPollData] --- g_hStateChangeEvent Received!
17:09:49 [D:GPS] GPS: [ProcessPollData] GPS data was received.
17:09:49 [D:GPS] GPS: [ProcessPollData] +++ Wait for g_hStateChangeEvent
17:09:50 [D:GPS] GPS: [ProcessPollData] --- g_hStateChangeEvent Received!
17:09:50 [D:GPS] GPS: [ProcessPollData] No GPS data was retrieved!
17:09:50 [D:GPS] GPS: [ProcessPollData] +++ Wait for g_hStateChangeEvent
17:09:50 [D:GPS] GPS: [ProcessPollData] --- g_hStateChangeEvent Received!
17:09:50 [D:GPS] GPS: [ProcessPollData] GPS data was received.
17:09:50 [D:GPS] GPS: [ProcessPollData] +++ Wait for g_hStateChangeEvent
You'll note that when polling twice per second, the gpsid ends up just hanging out on every other poll event, sucking up cycles for no particular reason because there's no data there to be had.

I haven't yet actually tested pollinterval 1000 and 100, but i can imagine at 1000 it'll be getting data every poll event, and at 100 there will be 9 poll events with no data and 1 with data.. but rather than conjecture, i'm off to test now :P

this would explain the Garmin switching back and forth from no sats to sats when set at 250....I'm gonna change from 500 to 1000 and test it out. I think the maxbuffer setting and input/output buffers helped reduce the lag. It would also probably explain the speed lag most are seing... Now if we can figure out how to gpsdriver to tell the radio to spit out 2x a sec or more, I think we'll reduce both speed and position lag! Excellent work ! Thanks.....getting closer..

[Da_G]

Sleuth255: not 1 sentence per second, one set of sentences per second..

when the gpsid goes to poll, it checks for new data and doesn't find any (because on a poll interface, the data is updated 'all at once' vs. streaming like on a COMM interface)

as for getting the gps unit to spit out more than 1 set of strings per second, i dont think it's gonna be possible, but who knows. at the least we'll need access to the NVRAM to flip some bits.. we've pretty much exhausted all configuration data that is read from the registry.

And yes, it is definitely the radio rom that generates all gps strings, I have been debugging that for a couple days now, which is why i say we have exhausted what we're going to do with registry settings.. AT commands and NV params are what we need to look at now.

While i've been at it I was debugging AGPS process to understand how it works. When GPSMode is set to 1 (MS Assisted) or 2 (MS Based), quickgps data is only used if agps data fails. process goes as follows:

gpsdriver issues at command to radio to poll cellid
radio returns cellid to gpsdriver
gpsdriver looks up following reg keys:
[/code]
[HKEY_LOCAL_MACHINE\SOFTWARE\HTC\SUPL AGPS]
"ServerIP" ; ip address, string
"ServerPort" port, string
[/code]
gpsdriver then sends your cellid info to this server (aka the PDE)
PDE server calculates estimated position based on cell id data and returns this info to gpsdriver
gpsdriver then looks at epheremis data from quickgps and guesses what it should be seeing in the sky from location/epheremis correlation

so people with nonfunctioning agps probably the issue is incorrect serverip/serverport settings, for example on my at&t rom the agps server is set to AT&T's agps server, which is only accessible from GPRS media net connection. so using my rom on another provider would require this to be changed to the proper pde server for that provider (or a public one, there are several, if u dont mind sharing your location data with them)
GPSMode set to 3, EnableAGPS to 0, and EnableReAiding to 0 is only true way to get fully standalone GPS operation - perhaps this should also be tested with the lag since every rom i have looked at has gpsmode set to 2.

[Sleuth255]

That will affect fix but not lag IMO. Once the chip is tracking birds assisted ephemeris data isn't needed since its getting it directly from overhead satellites.

[Da_G]

Depends on the mode and the registry settings.. with ReAiding enabled there's a periodic re-injection of assistance data.. with GPSMode set to 1 assistance data is used for every single fix.. but yeah, i agree not likely to affect the lag. just sharing the info since i learned it in the process of digging :)

[Sleuth255]

It is interesting stuff. I was under the impression that the aGPS server returns precise ephemeris data based on tower info. QuickGPS returns generalized ephemeris data. Both seem to be able to work independently... aGPS should result in consistent hot-starts. I wonder why this isn't the case on GSM but is the case on CDMA...

possible answer?

AT&T aGPS PDE = POS
Verizon aGPS PDE = works as advertised

[Da_G]

Also check out QOSAccuracy and QOSPerformance under SUPL AGPS.. if you set QOSPerformance to 0 you get no strings other than $PSTIS.. documentation i have for this is old and refers to it being one value that controls the amount of time chip spends searching for sats.. obviously not the case here.

[btprice2001]

Hi, Da_G. I just flashed your latest rom and I don't have most of these registry entries. Is that normal? Should I just use Sleuth's cab or add them? Specifically I don't have the following in my registry:

MaxLogFileSize
SleepOnNoData
OutputBufferSize
InputBufferSize
MaxBufferSize

[Da_G]

Yep it's normal, the entries aren't required for it to work, just tweaks to hopefully make it work better :) Add em manually or cab!

[noellenchris]

Also check out QOSAccuracy and QOSPerformance under SUPL AGPS.. if you set QOSPerformance to 0 you get no strings other than $PSTIS.. documentation i have for this is old and refers to it being one value that controls the amount of time chip spends searching for sats.. obviously not the case here.


Have you tested with QOSPerformance to 0? Do you think this would reduce the lag? Thanks...:)

[dizzyraider]

What's weird is that when I added those registries either manually or through cab, it breaks my Garmin XT. To get it working again, I had to remove these new registries. I'm running the 2.06 non MSVC version.

[mindfrost82]

What's weird is that when I added those registries either manually or through cab, it breaks my Garmin XT. To get it working again, I had to remove these new registries. I'm running the 2.06 non MSVC version.
What breaks on it? I had to set the PollInterval to 500 instead of 250.

[Da_G]

Sure, QOSAccuracy to 0 eliminates position lag entirely.

How you ask? It doesn't report a position at all. LOL

I think this value is used so that if the 'dilution of precision' is too high, it won't report a position at all, until the DOP is low enough to be under the threshhold.. that'd be my guess, but i haven't tested it much yet. Since I don't think that'll help our lag issue.

One thing I observed that might be a coincidence, is that when I got a fix in motion, i had pretty much zero position lag until i stopped at an intersection, and the 'static navigation'-esque feature kicks in that locks your position until you move xx meters away, then starts updating again.. it's not very common to get a fix in motion though.. (fast enough motion not to trigger static navigation)

[btprice2001]

Depends on the mode and the registry settings.. with ReAiding enabled there's a periodic re-injection of assistance data.. with GPSMode set to 1 assistance data is used for every single fix.. but yeah, i agree not likely to affect the lag. just sharing the info since i learned it in the process of digging :)

I've never been able to leave aGPS enabled without getting the problem where the GPS goes back and forth between having a fix and then seeing zero satellites every few seconds. I tried disabling ReAiding and enabling aGPS and so far I have not had any issues! Btw, I left GPSMode at 2. I don't know if my fixes are any faster , but it can't hurt, right?

[Da_G]

Yes, i'm seeing this behavior too.. in fact I don't think i've seen AGPS function as intended once yet.. this leads me to believe more and more that, like sleuth was saying, AT&T's PDE server is feeding us bogus data..

It might be worth our time to see if a public assistance server actually feeds correct data.. while it's not likely to help with the lag issue, being fed proper AGPS data would certainly help other functionality..

I'm tinkering now :)

[kk0813]

Iguidance 2009 was announced recently and was scheduled to start shiping from 01/05/2009. I am currently using Iguidance 3.01 and Copilot Live 7, both have old map data. I didn't find much change in Iguidance 4.0 so I skipped upgrading it. IGuidance 2009 (with 2008 Q4 map data) looks interesting, did any one order it yet?
I have speed lag in Iguidance more compared to copilot 7. If screen draw is causing any issue in terms of speed lag, according to IGuidance 2009 site, it uses eGML graphic engine for smooth lines, does it make any improvement in moving the map while driving to avoid speed lag?

[dizzyraider]

There is not problem when I change the PollInterval, but it is with the addition of the buffer registries that breaks my Garmin XT. When I fire up Garmin with the registries, it would open up for a couple of seconds then exits by itself. Then when i try to open it again, a small error window shows up and won't restart the program. It would work again after I take those registries out (soft reset after registry change).

[noellenchris]

Iguidance 2009 was announced recently and was scheduled to start shiping from 01/05/2009. I am currently using Iguidance 3.01 and Copilot Live 7, both have old map data. I didn't find much change in Iguidance 4.0 so I skipped upgrading it. IGuidance 2009 (with 2008 Q4 map data) looks interesting, did any one order it yet?
I have speed lag in Iguidance more compared to copilot 7. If screen draw is causing any issue in terms of speed lag, according to IGuidance 2009 site, it uses eGML graphic engine for smooth lines, does it make any improvement in moving the map while driving to avoid speed lag?

Here is what I found on google...sounds good I guess?

The eGML is a strongly speed optimised 2d/3d graphic library. The render engine is optimised to read out high quality images with smooth colour transitions and sleek bends. The eGML is basically designed to run on different hardware and software environments. PocketPC

OpenGL ES-compliant graphics library and embedded 2D/3D Graphic Multiplatform Library (eGML), for the real-time generation of display graphics; Imagem’s spline-based library for handling bitmaps, frame buffer accesses, drawing of graphic primitives and object handing and movement;



Sounds better....

[Da_G]

Raphael has native OpenGL ES drivers so that should indeed perform faster.

That being said, I would be hugely surprised if it made a difference in GPS lag..

No matter how terrible the performance of the graphics driver is, it should not cause lag, only frame dropping.. not as 'smooth' in other words..

So far my tests with AGPS are coming up with wierd results, it doesn't seem to be working at all. I'm thinking the RIL is set to use the IP/Port in NVRAM and not reading it out of the registry.. only way to tell is to set up a gprs packet sniffer though.. anyone know of one that works? :)

Tried the verizon AGPS server, supl.nokia.com, and at&t's, all are yielding the same results.. im thinking agpsnvsetting needs to be set to a value that ignores the NV agps servers and uses the registry ones.. assuming that's what that key controls. still debugging..

[yzf750]

I never seem to get a gdop value under 4 on my fuse. At least in the car. In the open yes its better (1.5-2.5). I contribute this as the reason to the position being off by 50-200 feet at times which is super annoying when you have 2 roads side by side within 50 feet or when it showing you 50 feet to the right of the road then jumping you on to off ramps at times passing them by. What are other people getting for readings in a car off the fuse gps?






DOP Value Rating Description
1 Ideal This is the highest possible confidence level to be used for applications demanding the highest possible precision at all times
2-3 Excellent At this confidence level, positional measurements are considered accurate enough to meet all but the most sensitive applications
4-6 Good Represents a level that marks the minimum appropriate for making business decisions. Positional measurements could be used to make reliable in-route navigation suggestions to the user
7-8 Moderate Positional measurements could be used for calculations, but the fix quality could still be improved. A more open view of the sky is recommended
9-20 Fair Represents a low confidence level. Positional measurements should be discarded or used only to indicate a very rough estimate of the current location
21-50 Poor At this level, measurements are inaccurate by as much as 300 metres with a 6 meter accurate device (50 dop * 6 meters) and should be discarded

[kk0813]

I never seem to get a gdop value under 4 on my fuse. At least in the car. In the open yes its better (1.5-2.5). I contribute this as the reason to the position being off by 50-200 feet at times which is super annoying when you have 2 roads side by side within 50 feet or when it showing you 50 feet to the right of the road then jumping you on to off ramps at times passing them by. What are other people getting for readings in a car off the fuse gps?






DOP Value Rating Description
1 Ideal This is the highest possible confidence level to be used for applications demanding the highest possible precision at all times
2-3 Excellent At this confidence level, positional measurements are considered accurate enough to meet all but the most sensitive applications
4-6 Good Represents a level that marks the minimum appropriate for making business decisions. Positional measurements could be used to make reliable in-route navigation suggestions to the user
7-8 Moderate Positional measurements could be used for calculations, but the fix quality could still be improved. A more open view of the sky is recommended
9-20 Fair Represents a low confidence level. Positional measurements should be discarded or used only to indicate a very rough estimate of the current location
21-50 Poor At this level, measurements are inaccurate by as much as 300 metres with a 6 meter accurate device (50 dop * 6 meters) and should be discarded

I just checked my fuze in home using GPS Test and HDOP is between 1.2 and 1.8 and some time going to 3.20 and immediately coming down to 1.4

[noellenchris]

I never seem to get a gdop value under 4 on my fuse. At least in the car. In the open yes its better (1.5-2.5). I contribute this as the reason to the position being off by 50-200 feet at times which is super annoying when you have 2 roads side by side within 50 feet or when it showing you 50 feet to the right of the road then jumping you on to off ramps at times passing them by. What are other people getting for readings in a car off the fuse gps?






DOP Value Rating Description
1 Ideal This is the highest possible confidence level to be used for applications demanding the highest possible precision at all times
2-3 Excellent At this confidence level, positional measurements are considered accurate enough to meet all but the most sensitive applications
4-6 Good Represents a level that marks the minimum appropriate for making business decisions. Positional measurements could be used to make reliable in-route navigation suggestions to the user
7-8 Moderate Positional measurements could be used for calculations, but the fix quality could still be improved. A more open view of the sky is recommended
9-20 Fair Represents a low confidence level. Positional measurements should be discarded or used only to indicate a very rough estimate of the current location
21-50 Poor At this level, measurements are inaccurate by as much as 300 metres with a 6 meter accurate device (50 dop * 6 meters) and should be discarded

Which program are u using to see GDOP? HTC test?
Which radio are u using?

[Da_G]

More info..

Using AGPSNVSetting = 5 causes the gpsdriver to load agps pde info from registry..


16:59:34 [D:GPS] GPS: [HTC_EnableConnManager] GPRSConnection = MEdia Net
16:59:34 [D:GPS] GPS: [NMEAThread] --- PosReqSentEvent RECEIVED!
16:59:34 [D:GPS] GPS: [SetupGPSConfigForGetPosition] Using AGPS settings from registry.
16:59:35 [D:GPS] GPS: [HTC_EnableConnManager] GPRS connection established!!
16:59:35 [D:GPS] GPS: [HTC_EnableConnManager] hHTC_Connection = 0x00000000, HTC_LocalIP = 1.1.1.1 (redacted)
16:59:37 [D:GPS] GPS: [SetupGPSConfigForGetPosition] Transferred from [supl.nokia.com] to [66.35.236.25]
16:59:37 [D:GPS] GPS: [SetupGPSConfigForGetPosition] NEW ServerIP = 66.35.236.25
16:59:37 [D:GPS] GPS: [SetupGPSConfigForGetPosition] AGPS Mode - ReAiding is Enabled.
Wimmob: ip address 0x19ec2342, ip port = 7275
pdsm: ip address 0x4223ec19, ip port = 27420
16:59:37 [D:GPS] GPS: [ProcessGpsFixRequest] ================================================== ==========
16:59:37 [D:GPS] GPS: [ProcessGpsFixRequest] operation_mode = 1
16:59:37 [D:GPS] GPS: [ProcessGpsFixRequest] num_fixes = 999999999
16:59:37 [D:GPS] GPS: [ProcessGpsFixRequest] time_between_fixes = 20
16:59:37 [D:GPS] GPS: [ProcessGpsFixRequest] accuracy_threshold = 50
16:59:37 [D:GPS] GPS: [ProcessGpsFixRequest] gps_session_timeout = 89
16:59:37 [D:GPS] GPS: [ProcessGpsFixRequest] server_adrs = 66.35.236.25
16:59:37 [D:GPS] GPS: [ProcessGpsFixRequest] port_id = 7275
16:59:37 [D:GPS] GPS: [ProcessGpsFixRequest] ================================================== ==========


In this test, i've set GPSMode to 1 to force agps results, and set to obtain a fix every 20 seconds to allow more time to the system to get a fix.. (GPSMode 2 falls back to standalone if AGPS has issues and default fix is once per second which is not possible on gpsmode 1 because assistance data must be transferred for every fix)

odd is pdsm's port reply, it doesn't match up with what i fed it..

note gps driver's reply of gps_session_timeout = 89 also, this matches up with registry setting QOSPerformance, older gpsone documentation talks about it..

anyway, thats what i found out for today :P

Oh yeah, and re: pollinterval, i've tested 100 500 and 1000 so far, and as i predicted at 100 there's 9 poll events with no data returned, 1 with return.. 500 is every other poll event, and 1000 data is returned on each poll event. so it seems there's nothing to be gained by polling faster other than wasting cpu cycles. :)

[Da_G]

Also observed the differing behavior of seeing 16 sats and narrowing it down vs. seeing one sat and building up from there, per the debug log the difference is if quickgps has downloaded and injected the epheremis data yet (injecting is not immediate after downloading)- if not, you start with 1 sat and build up from there.

[mindfrost82]

Have you looked into something like the PPST files for Sprint and Verizon?

I know when we were trying to get GPS working on the Sprint Touch we had an extremely difficult time until someone figured out we needed to run a PPST file that was in the /Windows directory to get aGPS to function. Someone guessed it wrote data into NVRAM or into the radio. Sure enough, once we got that file it worked great after that. The PPST file was different for each carrier too.

That could just be a CDMA thing though.

[Da_G]

believe me i wish it was that easy :) cdma raph test rom comes with all drivers etc necessary to access radio rom directly, gsm does not. and the tools for each are different (msm75xx vs. msm72xx)

i believe the method to access radio rom on our device is similar to method for ks20 (compositefn + windiag).. we have windiag but not compositefn drivers.. i have not been able to get ks20 ones working yet.. but i believe a working compositefn driver + usb switch is what we lack to get access to radio rom nv.

note that registereditor that i posted in raph software forum is capable of editing nv registers. but we need to know what address corresponds to what register..

[yzf750]

Which program are u using to see GDOP? HTC test?
Which radio are u using?

Stock att radio. I tryed 28,32 and m1 also. I think gps test which is floating around the forum shows the 3 dop values (v, h, and 3d) not sure as I use a few different programs at times. havent tried htc test yet. But to me its showing poor reception (unless totally outside of a car) which most likely is hardware related I think. My tilt was always under 2 in the car. I dont suffer a real lag issue, at least with igo8, maybe 1-1.5 sec most the time. Its the location (50-150) feet off that causes my problems. if i hold it out the window it gets much better 25-50 feet. it shouldnt be that bad in a car and its not just my car, it happens in any car ive tryed it in.

[Sleuth255]

VisualGPSce also shows DOP on the stats panel.

[noellenchris]

I was updating my "GeoScout" software for geocaching and under the sat info...it shows the data output and interprets it in a window below. Seems usefull, but I'm not Da_G. So it might as well be german to me.

[X2D]

if i hold it out the window it gets much better 25-50 feet.

I haven't had much opportunity to do much flashing and testing so I've stayed relatively quiet in this thread because I have nothing to contribute but when I read this comment this is what my face looked like --> :eek:

I flashed my phone within 3 days of having it but the thought of holding my phone outside a moving vehicles window is a disheartening thought lol.

You sir are a committed tester. ;)

[noellenchris]

I haven't had much opportunity to do much flashing and testing so I've stayed relatively quiet in this thread because I have nothing to contribute but when I read this comment this is what my face looked like --> :eek:

I flashed my phone within 3 days of having it but the thought of holding my phone outside a moving vehicles window is a disheartening thought lol.

You sir are a committed tester. ;)

Maybe I can try to scare my TP....especially if I hold it loosly out the window..lol:D

[kavi]

Another test:

The whole day I have been flashing different roms, radios (raphael) and applying patches mentioned all over.

I was testing the TP against my tilt running stock ATT and a Kaiser running HDIII Raphael. All the fixes so far just nudges the problem up and down but nothing really fixes the position lag issue. Lock times vary between ROM's and Radios but does not effect the position issue at all.

Results in brief:

Gradually Gaining speed:
REAL 20 30 40 50 60
KAIS 20 31 39 49 61
RAPH 15 27 35 44 59


Coming to a complete stop:
REAL 40 30 20 10 0
KIAS 40 31 21 8 3 2 1 0
RAPH 40 30 35 25 20 18 11 8 5 3 2 1 0


Passing cross street:
^ Direction of travel
==========================
----Actual Car
------------------Kaiser
==========================
------------------------------TP

Hope we can find an actual fix fast.

[Sleuth255]

Could you post a reg containing your GPS Intermediate Driver key and subkeys? What ROM/radio are you using now on your raphael? You should be able to tweak it to eliminate Position Lag with the verison of TT you are running.

[kavi]

Could you post a reg containing your GPS Intermediate Driver key and subkeys? What ROM/radio are you using now on your raphael? You should be able to tweak it to eliminate Position Lag with the verison of TT you are running.

I am using your regfix cab from the thread.
Radio 1.08.25.20M1 Raphael
ROM NATF 4.1 Standard

Reg entries are attached. It would be great if I can get this fixed.

[noellenchris]

Da_G, have you tried the latest xperia radio? Some ppl are saying it fixes the GPS....I haven't tried it yet. I'm guessing it's just luck....:confused:

[kavi]

Da_G, have you tried the latest xperia radio? Some ppl are saying it fixes the GPS....I haven't tried it yet. I'm guessing it's just luck....:confused:


Is it just faster lock times or a real fix for the pos. lag issue?

[noellenchris]

Is it just faster lock times or a real fix for the pos. lag issue?

In the radio thread people are saying it fixes the everything. But I've heard that before. I've got rid of most of my lag by reducing the maxbuffer to 256. But's it's still got speed lag. So it's a small achievement, but not a good fix. Da_G is working feverishly debugging the gps drivers and stuff beyond most of us. Hopefully he discovers the cause of these issues.

Also some ppl report that installing the AT&T Navigator fixes the gps too. I'm curious to see the added reg settings or drivers if any are added.

[kavi]

In the radio thread people are saying it fixes the everything. But I've heard that before. I've got rid of most of my lag by reducing the maxbuffer to 256. But's it's still got speed lag. So it's a small achievement, but not a good fix. Da_G is working feverishly debugging the gps drivers and stuff beyond most of us. Hopefully he discovers the cause of these issues.

Also some ppl report that installing the AT&T Navigator fixes the gps too. I'm curious to see the added reg settings or drivers if any are added.

I am also hoping that Da_G will put and end to this misery.

Did you try the ATT Nav App?

[noellenchris]

I loaded it, but didn't subscribe. I inspected the registry, and couldn't find any changes after install. All my nav programs work the same but haven't test lag. I was hoping to find some registry changes since it requested a reboot after install. I may activate it and see what it does....I'm also thinkin on changing to the xperia radio. This Blackstone radio is so good on the battery though...GPS time is soso. I think for me the M1 radio was quicker on fix...

[kavi]

I loaded it, but didn't subscribe. I inspected the registry, and couldn't find any changes after install. All my nav programs work the same but haven't test lag. I was hoping to find some registry changes since it requested a reboot after install. I may activate it and see what it does....I'm also thinkin on changing to the xperia radio. This Blackstone radio is so good on the battery though...GPS time is soso. I think for me the M1 radio was quicker on fix...

I will also try to install and see whether it makes any registry changes. Sadly reglogger is not working for me. Planning to do a whole registry compare with BC.

BTW I am using M1 and GPS lock is good although having band switch issues @ home that I never had with the Kaiser. :mad:

[Da_G]

I looked at the cab installer for telenav when i first dissected the fuze rom, it didn't seem to change anything gps related in the reg.. don't remember if i disassembled the exe to see if it does when it loads though..

BTW the new leaked sprint rom has new gps drivers, but i see they have some cdma specific code, let's see if they work/help.. :)

They're significantly different than the one's we're using.

[noellenchris]

I looked at the cab installer for telenav when i first dissected the fuze rom, it didn't seem to change anything gps related in the reg.. don't remember if i disassembled the exe to see if it does when it loads though..

BTW the new leaked sprint rom has new gps drivers, but i see they have some cdma specific code, let's see if they work/help.. :)

They're significantly different than the one's we're using.

was gettin worried...thought you found another thread to hang in...lol I'm flashing the xperia radio now. Would like to see the new drivers too. well you know more than me about what your lookin at. Is the Sprint TP's having any gps issues? Thanks again for all your work:D

[Da_G]

Duno, the ROM is so new I don't know if anyone's running it yet..

But with the new driver, no response from GPS unit at all, probably the CDMA specific code at fault there.. doh!

[kavi]

I looked at the cab installer for telenav when i first dissected the fuze rom, it didn't seem to change anything gps related in the reg.. don't remember if i disassembled the exe to see if it does when it loads though..

I guess then it's pointless trying to capture the registry changes. Will wait for your findings.

Thanks for all the hard work Da_G:)

[Sleuth255]

I am using your regfix cab from the thread.
Radio 1.08.25.20M1 Raphael
ROM NATF 4.1 Standard

Reg entries are attached. It would be great if I can get this fixed.

Two Questions:
Do you have the option for a "Builtin GPS receiver" device in TomTom?
Have you tried the 1.02.25.28 Radio?

[mindfrost82]

I loaded the Xperia radio last night and tested GPS this morning.

The initial fix was pretty fast, probably faster than the Blackstone one.

The lag was still there but maybe a little less than it was with the BS radio.

When going 40mph using iGo, it was like the point of the arrow was at the back of my truck rather than the front. In other words, when the back of my truck crossed the intersection is when the point of the arrow would cross the intersection. So that's like 20ft off and probably less than 1 sec lag.

I'll have to test if the call quality is on par with the BS radio though. Call quality and reception is #1 for me and GPS is after that. The Xperia radios have always given me the best reception but the call quality hasn't been as good as the BS radios.

[noellenchris]

I loaded the Xperia radio last night and tested GPS this morning.

The initial fix was pretty fast, probably faster than the Blackstone one.

The lag was still there but maybe a little less than it was with the BS radio.

When going 40mph using iGo, it was like the point of the arrow was at the back of my truck rather than the front. In other words, when the back of my truck crossed the intersection is when the point of the arrow would cross the intersection. So that's like 20ft off and probably less than 1 sec lag.

I'll have to test if the call quality is on par with the BS radio though. Call quality and reception is #1 for me and GPS is after that. The Xperia radios have always given me the best reception but the call quality hasn't been as good as the BS radios.


LOL.....I did the same thing last night and my results are just like yours. I did make some calls today and it seems just as good as the BS. So far the batter seems just as good also. The GPS does seem faster than BS (For me anyway). The internet seems good to go also. Lag seems the same too.

I did notice though that "Nan" is still there.
My Garmin was flashing back and forth from one red dot to full bar in gps signal strength with the BS radio, and now it is steady with good strength. But it still lags identically to the BS. TT7 seems the same also. The speed lag did seem a tiny less (could be due to the threat's I made about hanging it loosly out the window...lol).

Sleuth: did you ever figure out that nan thing? It always seems to pause at that $PSTSa? :61 string. :confused:

Mindfrost: does iGo give traffic etc.? like Garmin?

[Sleuth255]

NaN doesn't affect anything since it's only an initial value that goes away as soon as you start moving. I don't think we'll ever change speed lag because its a function of the chip which I've also seen expressed in the actual $GPVTG sentences (when stopped, I've watched $GPVTG sentences go by, dropping the speed to zero at the same rate). $PSTIS is a proprietary sentence not used by navigation software. I thought Da_G mentions it in this thread as an input sentence. I think the pause is simply the 1 second poll rate. You can see this in HTC GPS tool. Sentences seem to pulse once per second. Sometimes $PSTIS is the last sentence before the next poll and sometimes it isn't.

[mindfrost82]

No, iGo doesn't do traffic unless you can do the TMC thing over the FM transmitter I think.

That's one thing I liked about Garmin, well that and the real-time Fuel Prices.

I just wish it had TTS and smoother map updates.

Oh, the other thing I noticed is in VisualGPSCe, my 3 DOP readings were around:
1.6-1.9
0.6
1.6-1.9

That was all the time while driving about 40mph. I never looked to see what it was at with the BS radio.

[mindfrost82]

Oh yeah, these are the settings I've been using. Since you guys said there's basically no reason to put the PollInterval below 1000 (because it only updates once per second), I changed that to 1000 and I also increased the SleepOnNoData to 1500 since Garmin seems to want that higher than the PollInterval, otherwise the signal keeps jumping around.

HKLM\System\CurrentControlSet\GPS Intermediate Driver\Drivers\GPSOneDevice
PollInterval = 1000

HKLM\System\CurrentControlSet\GPS Intermediate Driver\Drivers
InputBufferSize = 256
OutputBufferSize = 128
SleepOnNoData = 1500

HKLM\SYSTEM\CurrentControlSet\GPS Intermediate Driver\Multiplexer
MaxBufferSize = 256

HKLM\System\CurrentControlSet\GPS Intermediate Driver
MaxLogFileSize = 0
CurrentLogFile = null
OldLogFile = null

[noellenchris]

Oh yeah, these are the settings I've been using. Since you guys said there's basically no reason to put the PollInterval below 1000 (because it only updates once per second), I changed that to 1000 and I also increased the SleepOnNoData to 1500 since Garmin seems to want that higher than the PollInterval, otherwise the signal keeps jumping around.

HKLM\System\CurrentControlSet\GPS Intermediate Driver\Drivers
PollInterval = 1000

HKLM\System\CurrentControlSet\GPS Intermediate Driver\Drivers
InputBufferSize = 256
OutputBufferSize = 128
SleepOnNoData = 1500

HKLM\SYSTEM\CurrentControlSet\GPS Intermediate Driver\Multiplexer
MaxBufferSize = 256

HKLM\System\CurrentControlSet\GPS Intermediate Driver
MaxLogFileSize = 0
CurrentLogFile = null
OldLogFile = null

Interesting observation with the sleep to 1500. It makes sense though, you wouldn't want the GPS trying to go to sleep.. Otherwise my settings are identical. I'm gonna change my sleep to 1500 too.

I'm still trying to see if I can change my dpi (so i can run in QVGA) so that GuitarHero will work, and also Garmin may perform better. Strange course of action, but I'm really curious... :D

[iRiKi]

When going 40mph using iGo, it was like the point of the arrow was at the back of my truck rather than the front. In other words, when the back of my truck crossed the intersection is when the point of the arrow would cross the intersection.

intersections arent a good way of measure IMHO.

one of the most fiable test is driving in a highway, and when passing under a bridge check the gps and you'll see how much lag your device has :)

[noellenchris]

I'm going back to the BS radio. I had an echo on 3 different calls today....:(

[kavi]

I have finally decided to go back to my faithful Kaiser. I've had enough playing with the GPS. I have been using the Kaiser in the evening and now I know what I've been missing :)

More issues I noticed with TT7 and TP:

> Loss of signal at the end of a voice call (using BT). Takes a few seconds to regain location and if you are just about to make a few corners it really messes things up.
> Not every warm start is the same all though they are at even intervals.

Also the band switch delay is longer than the Kaiser and the time without signal is considerably higher. It maybe just my device but I seriously expected more from HTC. With time and this much user feedback they should be able to deliver on basic functionality.

I am thinking of going for the Xperia since the GPS lock and slide and type is just amazing. I will continue to keep an eye on the developments hopefully Da_G will come up with something.

Thank you all for the great tips and help.

[iRiKi]

I am thinking of going for the Xperia since the GPS lock and slide and type is just amazing.

It's almost the same as the Fuze:

Qualcomm MSM7200A Embedded gpsOne GPS module

Find the Xperia zone in the forum, they're having the same problems as we do :)

[kavi]

It's almost the same as the Fuze:

Qualcomm MSM7200A Embedded gpsOne GPS module

Find the Xperia zone in the forum, they're having the same problems as we do :)

That's news need to read more then. But my friend's one just works flawlessly with TT7. Thanks for the heads up. Looks like it's kaiser for me for a while. :rolleyes:

Edit:

Some more things I noticed while doing my tests (hope this helps the the testers)

> The driving speed on GPSTEST is quite accurate (almost the same as Kaiser) and in sync with the vehicle speed
> HTC GPS tool reports the laggy speed in sync with the TT speed
Most important
> The voice prompt is dead on. How? Well the same route, same map and same version of TT running on Kaiser and TP will announce at the same time. This leads me to believe that the display drawing is the one that's behind.Of course I could be wrong :)

My theory:
TT will announce the next turn depending on the distance from turning point. This is set by the program and it's a constant. If both devices are announcing at the same time that means both TT's know exactly how far you are from the next turning point (GPS cordinates must be identical or very close). Why it displays the position differently is the million dollar question!!!

[Sleuth255]

I'd at least flash Da_G and the blackstone radio before giving up on your fuze. Unless you choose not to believe your eyes, the comparison video on post 400 clearly shows that the problem can be solved. It documents the fact that I have zero lag and better signal strength than my tilt. The GPS on my fuze is superior IMHO and I'm running the same version of TT that you are.

Hint: Da_G has paid special attention to graphics performance. I'd try the exact 2.04-MSVC version I'm using first.

[mindfrost82]

I'm going back to the BS radio. I had an echo on 3 different calls today....:(
I also went back to the BS radio. This Xperia radio was just like the others I've tried in terms of call quality.

[kavi]

I'd at least flash Da_G and the blackstone radio before giving up on your fuze. Unless you choose not to believe your eyes, the comparison video on post 400 clearly shows that the problem can be solved. It documents the fact that I have zero lag and better signal strength than my tilt. The GPS on my fuze is superior IMHO and I'm running the same version of TT that you are.

Hint: Da_G has paid special attention to graphics performance. I'd try the exact 2.04-MSVC version I'm using first.

Ok then it's time to CID unlock and try what you are suggesting over the weekend.

BTW did you notice the drop of signal at the end of a call with the combo you are using?

[Ouzo]

kavi, please could you let us know how you get on with Da_G's ROM? So far I've changed the radio on my Touch Pro, but I haven't plucked up the courage to try a different ROM yet..... took me a whole year to flash my old Elf, and it completely transformed it!

I've tried TomTom 6.0, 6.030, and now 7.910(?) with UK maps 8.15 and no amount of reg hacking has helped me get rid of the lag. Like you, the voice is almost always spot on in TomTom so I too had begun to wonder if this is a video issue.... the graphics performance on my Pro is worse than on my Elf in certain instances (I know some people don't have lag).

[Sleuth255]

Ok then it's time to CID unlock and try what you are suggesting over the weekend.

BTW did you notice the drop of signal at the end of a call with the combo you are using?

I've never had a call come in while using the GPS. Sorry.

Read through this (http://forum.xda-developers.com/showthread.php?t=448008) thread. Leaving the SIM/SD cards out is a bit excessive and completely unnecessary IMHO but the rest is good stuff. You need to get to hardspl before doing anything. No CID unlocking is necessary when its installed. Blackstone radio is your last test and may be unnecessary. It will cost you to go there anyway since you need to security unlock to run it.

[Da_G]

More info: if you have both ServerURL and ServerIP under \Software\HTC\SUPL AGPS, it pulls the value from ServerURL and uses that. (not related to lag, just sharing more info i found in the process of poking)

Another observation: The NMEA thread reports a Priority Level (http://msdn.microsoft.com/en-us/library/aa450618.aspx) of 251, which is a non-real-time priority level.. this could be allowing anything using more cpu with a higher priority to take precedence, causing/adding to the lag issue..

[Da_G]

Note the times in this log: (i'm replying to myself to bump thread because this seems significant)


18:13:38 [D:GPS] GPS: [ProcessPollData] --- g_hStateChangeEvent Received!
========POSITION REPORT========
valid 0Xfffdfyear = 2009, mon = 1, day = 18, h = 2, min = 13, sec = 37, ms = 200


There is a consistent one second lag in the position report vs. the log timestamp (log timestamp generated from time on cell towers, which should be getting its timesource from the same place as the gps chip, the satellites.. so that shows that there is a real 1 second lag somewhere in the system before the NMEA data even hits the GPSID..)

These values are generated directly by gpsdriver.dll so that shows gpsid_htc.dll is definitely NOT at fault with the gps lag, and it's either gpsdriver.dll or even further up the system in the radio rom itself.. i'm still thinking it's a simple NVRAM bit flag..

[noellenchris]

More info: if you have both ServerURL and ServerIP under \Software\HTC\SUPL AGPS, it pulls the value from ServerURL and uses that. (not related to lag, just sharing more info i found in the process of poking)

Another observation: The NMEA thread reports a Priority Level (http://msdn.microsoft.com/en-us/library/aa450618.aspx) of 251, which is a non-real-time priority level.. this could be allowing anything using more cpu with a higher priority to take precedence, causing/adding to the lag issue..

Can we change Priority Level for the GPS via reg?

[Da_G]

You're supposed to be able to change it by setting the key:


[HKLM\Drivers\BuiltIn\DriverName]
"Priority256"=dword:0


But it doesn't look like the gps driver reads that key, as it reports the same priority afterward.. looking into it :)

[Da_G]

While i'm digging around in here, gpsdriver.dll has a hardcoded fallback agps pde server of supl.agps.emome.net, maybe we can try using that and see if it feeds us valid data.. it does return to a valid ip.. haven't checked if the service is running though.

[noellenchris]

You're supposed to be able to change it by setting the key:


[HKLM\Drivers\BuiltIn\DriverName]
"Priority256"=dword:0
But it doesn't look like the gps driver reads that key, as it reports the same priority afterward.. looking into it :)


Can we add a priority to HKLM\System\CurrentControlSet\GPS Intermediate Driver\Drivers etc.

The more I dig into the registry, the more questions I have....
There are alot of settings for the SDHC etc...I wonder if we could speed that up too. ? I'll just have to save that for another project. lol

[Da_G]

Well, sure you could put the key there, but the program doesnt read it :(

So far I haven't seen any priority key that it reads.. :(

[Black1982]

If both devices are announcing at the same time that means both TT's know exactly how far you are from the next turning point (GPS cordinates must be identical or very close). Why it displays the position differently is the million dollar question!!!

This (the audio cue being spot on, while the display is lagging)has been reported many times. I also have this experience

My simple logic: tomtom simply can not update screen fast enough!
I can only guess at what is causing this.
(maybe something to do with the notorious drivers issue)

Posts bij Da_G, Sleuth ea seem to indicate that there is nothing wrong with
the GSP system and settings (at least to me).

Is there any solid method of testing this hypothesis?

[Sleuth255]

The voice synchronization reports are what let me to speculate that graphic drivers are behind (at least some of) the lag issues. I need to stress "at least some of" here because I really think there's more than one position lag root cause.

Key to this is going to be folks like Kavi that load Da_G's ROM and report. Using v2.04-MSVC as a baseline for this test is also important as it will reproduce my exact configuration from a loaded ROM standpoint. Tests should be conducted with TomTom using the base ROM (disable UC by pulling your SD card) and with the GPSIDRegFix cab applied. Anyone who has noted voice synchronization should try this.

[Ouzo]

Well, as my audio is in the right place but the image is not, I guess this makes me a candidate to try your suggestion Sleuth255.

Noob question, but will TomTom Home have a fit if I flash a new radio (I'm only on 1.02.25.28) and ROM to my device? I don't really want to have to pay for the software a second time.

Also Sleuth, my 8gig SD card is only a class 2. Would that screw up the comparison?